India - Irrigation sector review (Vol. 1 of 2) : Main report

Report No. 9518-IN India Irrigation Sector Review (in Two Volumes) Volume l: Main Report December 20, 1991 Agriculture Olperations Division MICROFICHE COPY India Country Department AsdiaRegiountryepartmenReport No. 9518-IN Type: (SEC) Asia Region OBLITAS, KI X80286 / / AS4AG FOR OFFICIAL USE ONLY zN Y ,c~~~~~~~~~~~C Docment of the World Wank '4't '', This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otlherwise be disclosed without World Bank authorization.  INDIA IRRIGATION SECr-OR REVIEW Abbreviations AAR Annual Administration Review AD Agricultural Department CADA Command Area Development Authority CBIP Central Board of Irigation and Power CGWB Central Groundwater Board CWC Central Water Commission ERR Economic Rate of Return GOI Govemment of India ID Irrigation Department MIS Management InfoTrmation System MOA Ministry of Agriculture MOF Ministry of Finance MOWR Ministry of Water Resources NABARD National Bank for Agriculture and Rural Development NGO Non-Governmental Organization NWB Nadonal Water Board NWDA National Water Development Agency NWMP National Water Management Project O&M Operations and Maintenance PCR Project Completion Report R&R Resettlement and Rehabilitation UNDP United Nations Development Program WALMI Water and Land Management Institute WYJA Water Users Association Exchange Rate: 17.90 Rs/US$1.00 (1990/91) Mem re: 1 hectare (ha) = 2.47 acres The Indian fiscal year is April 1 - March 31.  FOR OFFICIAL USE ONLY TITLE INDIA: IRRIGATION SECTOR REVIEW COUNTRY INDIA REGION SOUTH ASIA SECrOR : AGRICULTURE (IRRIGATION) Rort P CLASSIFICATION MMYY LANGUAG 9518-IN SRA Restricted 12191 English ABSTRACT India's achievement in rapidly expanding its irrigated area, now the largest in the world, underpins nearly two-thirds of the country's agricultural production, but yields have been disappointing, limiting agricultural growth. Sector management has deteriorated and the sector now faces financial crisis and deteriorating infrastructure. Improved sector performance will be critical to India's agricultural growth. This report makes a diagnostic of the current issues and opportunities in the Indian irrigation sectoi and on this basis provides a proposed strategy and action plan for the Indian government, at both central and state-levels, for irrigation developrment through the 1990s. The recommended strategy focusses on four action areas. First, to develop efficiently the country's water resources, there is a need to forge and implement a coherent water resources policy based on river basin planning and strengthened data collection and coordination between users and states. Second, government expenditures need prioritization, emphasizing higher performance of existing irrigation and the most viable expansion prospects, and financial management of expenditure and revenues needs to be tightened. Third, technical perfornance needs upgrading, particularly in the areas of ensuring adequate maintenance, better water management, good quality of construction, improving environmental impact and ensuring full rehabilitation of communities displaced by irrigation development. Finally, at the heart of these reforns is the need to restructure irrigation institutions, enhancing their management and technical capabilities and progressively devolving responsibilities to farmers and other non-government entities. This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  INTRODUCT1f Q The objective of this report is to provide a diagnostic of current issues and opportunities in the Indian irrigation sector and on this basis, to propose a strategy for the Indian Government at both central and state-levels for irrigation development through the 19('. The report has Leen produced in collaboration with the Indian Government and is based on field work and analysis, and discussions between Government and the World Bank, in the 1990/91 period. Further discussions of the draft report ("Green Cover" dated June 27, 1991) were held with Govemment in September/October 1991. These included regionally based meetings with state Irrigation Departments, and discussions with the Ministry of Water Resources and its member agencies, and with the Planning Commission, Ministry of Finance, Ministry of Agriculture and Ministry of Environment and Forests. This final version of the report has incorporated revisions resulting from these discussions. The report is in two volumes. This volume, the "Main Report", contains the full strategy and recommendations resulting from the Irrigation Sector Review. Volume II, entitled "Supplementary Analysis", contains additional discussion and analysis in related areas as back-up material for Volume I. Acknowledgements. Report preparation has been handled by the Agriculture Operations Division, India Country Department, with contributions from staff in the Asia Technical Department, Agriculture and Rurai Development Department, the New Delhi office and government representatives. From the Bank's side, K. Oblitas was task manager and principal author with input and review contributions from B. Albinson, W. Barber (Cons.), M. Baxter, J. Berkoff, R. Bhatia (Cons.), B. Bhavanishankar, S. Blinco (Cons.), T. Blinkhorn, A. Bottrall (Ford Foundation), V. Byrappa (Cons.), D. K. Dutt (Cons.), C. des Bouvrie, P. Duane, H. Frederiksen, Ci. Fauss, J. Gianchandani (Cons.), G. Kathpalia (Cons.), K. Kadiresan, S. Khan, R.K. Malhotra, M. Mitra, A.A. Pai, R. Palomares, W. Partridge, J. Pike (Cons.), C.C. Patel (Cons.), C. Perry, W. Price, V. Ravishankar, C. Rees, S. Singh (Cons.), L. Smedema (Cons.), A. Salam, J. Talbot (Cons.), and R. Venkataraman. Report processing was handled by B. Scott, C. Pierpont, C. Gan, and T. Sharrna. Principal reviewers were P. Cox, D. Greene, J. Wijnand, J- P Baudelaire, C. Diewald and H. Plusquellec. The assistance and advice provided by the Ministry of Water Resources (including CWC and CGWB), the Ministry of Finance, the Planning Commission, the Ministry of Agriculture and the Ministry of Environment and Forests, and not least, by the governments of the states visited, has been most valuable and is gratefully acknowledged. The report is also indebted to the advice received from the Indian academic community. Notwithstanding the invaluable help received, responsibility for the report's findings and recommendations is under the World Bank's India Country Department.  INDIA pSIEGAINm S,ECrO TOEEIEW Volume I - Main Report Table gConet Executive S u m m a ry ........ ................................. i-v I. THE CHALLENGE AND STRATEGIC CHOICES ................. 1 A. Critical Issues .......... 1 Productivity of Irrigation .1 Sustainability .1 Investment Focus and Financial Disipline .1 Ineffective Sector Management .......................1... B. Development of Irtigation .... ........................... 2 Agroclimate ................................. 3 Regional Socioeconocs.. 3 Diversity Between States. 3 C. Inpact of Irrigation. 4 Cropping Intensity and Expansion of Cropped Area. 4 Crop Yields. 4 Output Stablization. 4 Crop Diversification. 4 Farm Incomes and Employment. 4 Secondary Income and Employment Effects. 5 Poverty Alleviation .5 Regional Development. 5 D. Potential for Irrigation Development ......................... 6 Irrigation and Drainage ............................... 6 Regional Development. 8 E. Economic Viability of Irrigation and Drainage Investment. 9 Past Performance and Viablity. 9 Returns by Type of Irrigation Investment .10 F. Needs for Future Investment . 12 Strategy Shift .12 Investment Choices .12 G. Framework for Strategic Choice .14 Forge a Coherent Water Resources Policy ................... 14 Prioritize Expenditure and Tighten Financial Management .14 Improve Technical Performance .14 Reorder Govemment/Private Roles and Build Capacity .. ...... 14 II. WATER RESOURCE POLICY AND PLANNING .................. 15 A. Situation and Trends .................................. 15  Deo -.opment Stmtere .g.y........... . .. .. .. .. .. .. .. . . 15 Linited Availabillity ................................ 15 Competing Uses of Water ......... ................... 16 Data Collection and S h aring 16 Institutional Constraints and Mechanisms 16 Water-Sharing and River Basin Planning .17 Watershed Developent .17 Coordination Needs .18 B. Tackling Resource Planning and Policy Issues .18 River Basin Planning .18 Inter-State Cooperation .18 Cenal Coordination. .18 Water-Picing .19 C. Actions to Improve Water Policy and Planning .19 III. EXPENDITURE AND FINANCIAL MANAGEMENT .21 A. Investment Focus ............. 21 Situation and Trends ................................ 21 Expenditure Priorities ............................... 21 Project Preparation and Selection ........................ 23 Actions to Improve Project Investment ..................... 24 B. Financial Management ................................. 24 Rising Current Expenditures ............... .. ..... 25 Inadequate Funding of Maintenance ....................... 25 Increasing Investment Costs ........................... 26 Declining A -venues from Irigation ....................... 26 Actions t- rove Financial Management ................... 26 Cost Rec-. very Policy and Procedures ..................... 28 Actions to Improve Cost Recovery .29 C. Funding Conditions by Central Governent .31 Efficiency, Maintenance and Cost Recovery .31 IV. TECHNICAL PERFORMANCE ..33 A. Productivity .33 Water Management .................... ............. 33 Water Distribution Systems ........................... 33 Water Management by Region .......................... 34 Water Management, Research and Extension Issues ............. 36 Acdons to Improve Water Management .............. ....... 38 The National Water Management Project .. ............... 38 B. Sustainability ....................................... 39 Design ........................................ 40 Construction Quality ................................ 40 Actions to Improve Construction Quality .................... 41 Maintenance ..................................... 42 Actions to Improve Maintenance ......................... 43 Environmental Issues and Trends ........................ 44  Tack!ing Environniental and Resettlement Issues ............... 46 Actions to Improve Environmental Assessment and Impact ......... 46 Actions to Improve Resettlement &Rehabilitation ............... 48 V. GOVERNMENT/NON-GOVERNMENT ROLES AND CAPACITY .... ... 50 A. Situation and Trends .................................. 50 Doninant Public Sector .............................. 50 State Itrigation Departnents ............................ 51 Coordination ..................................... 52 Farmer Involvement ................................ 53 B. Using tne Non-Government Sector ......................... 54 Private Investmeiit Opportunities ........................ 54 Creating Autonomous and Accountable Institutions .... ......... 55 C. Improving Government Insdtutional Perfornance ......... ...... 57 Restructuring Irrigation Departments .57 Coordinating Government Institutions .59 Increased Role for Farmers and Women .................... 60 ANiNEX I Matrix of Highlights of Recommended Action ANNEX II List of Background Papers ANNEX III Bibliography List of Tables in the Main Text Table 1.1 Development of Irrigation: 1950-1985. 2 Table 1.2 Government Esfimates of Status of Indian Irrigation Development in 1985 7 Table 3.1 Revenues and Expenditures on Major & Medium Irigation Projects ... 25  EXECUTIVE SUMMARY i. In many respects Indian irrigation is at a crossroads. The choices are continuation of the status quo and a modest, possibly diminishing agricultural growth rate, or concerted reformn and renewal in the interest of matching reality to potential. The reality is that one of the world's largest irrigation investments is performing unevenly and, on average, far below potential. The yield gap is apparent within India -- betvieen the best-performing areas and those weighing down average yields -- and between India and other Asian countries. Cropping intensity also falls far short of performance in comparable settings in Asia and elsewhere. Despite several decades of significant public investment in irrigation infrastructure, mainly large surface schemes, productivity is not breaking through to enable agricultural growth beyond the past threshold of 2.5% per annum. ii. India will be critically dependent on better performance from irrigation. Over 55% of agricultural output is from irrigated lands, and production elsewhere is constrained by lack of land for expansion and the risks and limitations prevalent under rainfed conditions. The pricing environment, other rural infrastructure and rural services are also fundamental to agricultural growth, but the resolution of water constraints enabled by irrigatdon is a springboard for productive capability. Where irrigation has been successful in India, notably in the northwest, it has had wide- reaching impact on both agricultural and general welfare and development. By contrast, the commoner situation of wealdy performing irrigation has not allowed rural development potential to be realized. -ii. The findirngs of this review point to poor sector planning and financial management, on the one hand, and inadequate water management and maintenance, on the other, as the main causes of mediocre performance. Paradoxically, India's major engineering achievements over the past half- century, resulting in massive expansion of surface irrigation, have contributed to the sector's current problems. With the focus centered on construction, the broader management needs of the sector were neglected, and the cumulative costs of this neglect are now apparent. Over the past decade, the situation appears to have worsened: lack of financial discipline and accountability, neglectece maintenance, and construction abuses have become endemic. Also, rapidly rising recurrent expenditure on unproductive staff costs is squjeezing out genuine investment in the sector and, combined with inadequate pricing of water and electricity, is throwing the sector into financial crisis. India cannot afford an over-expensed and under-performing sector. Sooner, rather than later, the burden will be financially unsustainable, and infrastructure will be physically unsustainable due to declining construction and maintenance standards. The situation is compounded in some areas by environmental degradation. Above all, agricultural growth will suffer. iv. The role of management in creating as well as disentangling the knot of diagnosed problems in this review is clear. Much of the current sectoral malaise stems from the fact that irrigation is largely managed by government monopoly, that a culture of "government needs to do it" prevails, and that tlke sector's bureaucracy has grown unwieldy, not adaptive to changing needs, with narrow engineering interests and lacking training and incentives to improve performance. v. The most fundamental order of business is for government at both state and central levels to reassess its monopoly and to scrutinize its role vis A vis the non-government sector, whose tremendous energies and capital -- both hurnan and financial -- have barely been harnessed. Where appropriate, and opportunities abound, government should begin to divest as much implementation and investment as possible to the non-government sector. vi. Simultaneously, government should evaluate its own performar.ce, implement improvements in critically deficient functions, and ascribe new roles as nece,Osary to state and central institutions. This stock-taking exercise can serve the useful prrpose of charting a course for remediation and development and identifying short- and medium-term steps to take in getting there.  ii vii. Major effort is required in four areas that are the focus of this sector review: (i) forging a coherent water policy, (ii) prioritizing investment and getting control of expenditure, (iii) improving productivity and ensuring sustainability, and (iv) building critical capacity within the public and private sectors in order to manage the sector more efficiently and effectively. (See Highlights of Recommended Action at Annex I). viii. Water Policy and Planning. The present situation whereby planning of water development is handled disparately, with limited coordination between states and sub-sectoral users, skirts a primary requirement: that planning and coordination of water use occur on a holistic basis, centered on the natural hydrological unit, the river basin. Only one of India's 18 major river basins is within the confines of a single state. Further, demands for water -- whether for agricultural, municipal or industrial use -- are growing, and interests between different users and states are increasingly colliding. Data collection is also disparate and provides an inadequate basis for water planning. The magnitude of India's water resources, how they should be best deployed, to what areas, for what uses, are critical questions that deserve considered answers beyond the confmnes of individual users and state boundaries. These problems heighten the need for application ot india's National Water Policy, initiated in 1987 but with litt!e operational follow-up. The states should ratify the national water poli -y and develop state water policies fitting within the context of this document; river basin planning commissions should be estaolished to plan and develop water resources on a holistic basis; data collection should be strengthened; Pnd actions taken to improve coordination between states, users within states and with the center. ix. Investment Focus and Financial Management. Efficient use of scarce public resources must govern future investment choice and expeiiditue. Weil-conceived investment priorities and quality project preparation are essential. The investment focus for irrigation should shift toward actions that have high economic viability, near-term impact and relatively low per-hectare cost, and to those that support investment by the private sector. Past investment in high-cost construction of new surface schemes has led to a substantiai backlog of incomplete works that have absorbed funds and managerial effort to the detriment of more viable options and maintenance of existing schemes. Investment in new construction should be restricted to the few cases where regional development is dependent on the supply of additional water. x. In nearly all states, the development agenda through at least a decade should concentrate on improving the performance of existing irrigation, complemented by selective viable expansion. Achieving better performance demands that major attention be given to maintenance and to improving water management. Maintenance should have the first call on public expenditure -- preserving existing infrastructure is paramount -- followed by concerted effort and resource allocation to improve water management. Expansion, for the public sector, should involve completion of ongoing projects, with low remaining costs and rapid maturity, selected on the basis of clearcut viability, available water and sound technical preparation. Another opportunity is surface drainage, widely neglected in the past, in areas subject to monsoon waterlogging. Rehabilitation of failed infrastructure, and in some areas, selective investment in flood control, drainage to combat rising groundwater tables, and rehabilitation of saline lands are also needed. xi. While public investment in irrigation is the focus of tLis review, two important investments should be divested to the private sector groundwater and farm-level watercourses. Groundwater development, having the largest impact in recent decades both on expansion and productivity, should remain in the hands of farmers. Existing public tubewell development programs, which have .t been effective, should be halted, and existing wells turned over to farmers or entrepreneurs. The public sector's role should be to foster private investment: provision of credit, research and extension, monitoring of groundwater levels and planning of surface irrigation for conjunctive use and groundwater recharge. Farm-level watercourses, field channels and field drains (micronetworks) should be operated and maintained by farmers, who should also play a  i.i larger role in construction. Government implementation, which has been costly and with limited impact, should be replaced by technical assistance and credit to farmers. xii. Soaring recurrent expenditures -- increasing nearly threefold in real terms over the past decade -- must be curbed through halting the recruitment of staff and tightening financial controls against leakages. Steps must also be taken to bolster cost-recovery in the sector; water and electricity rates shoufii be treated as utility fees with rates govemed by O&M costs at minimum. To stimulate action and achieve accountability, water charges should be reassessed by state governments and cost-recovery performance evaluated annually in public reports. Collection should be simplified and handed over to local communities, retaining a percentage as incentive. As users take over scheme management, revenues collected should be retained. Electricity subsidies for well irrigation should be removed and meter-based pricing introduced. xiii. The skills that imbue financial discipline -- cost accounting, monitoring, financial and economic analysis, and business planning capabilities -- are sorely lacking in both state and central institutions. These skills should be implanted as quickly as possible, whether through training programs or recruitment of consultants. xiv. Accountability and transparency of the funding and expenditure process are critical requirements. Getting there will require state and central govemments to monitor more carefully financial management by state irrigation departments, based on annual analysis of expenditure, detailed planning and budgeting of maintenance and reporting of cost recovery status. Disbursement of central funds for state irrigation programs should be conditional on adequate financial and cost recovery reporting. control of recurrent expenditures, and full funding of maintenance. xv. Technical Performance. Improving water management should become the principal technical preoccupation of irrigation managers. Deficiencies are widespread, with some 10 -13 million hectares of surface irrigation performing well below potential and some improvement possible on most of the remaining area. Pilot efforts, including the National Water Management Project (Cr. 1770), are in their early stages but indicate the direct ons to take. The dimensions of the problem require a low-cost, replicable approach with emphasis on diagnostics of problems and improvement possibilities at the level of each command, followed by development, in consultation with farmers, of a seasonal plan of command operation, complementary investment in low-cost infrastructure to achieve better water control, and upgraded daily management and monitoring. The existing pilot efforts should be expanded, necessitating funding for this purpose, reallocation of management efforts, staff and training to this function, and use of consultants to augment current capabilities. Technological research should support this drive, also to be accompanied by piloting of innovations on selected commands to achieve higher productivity. xvi. Environmental and resettlement and rehabilitation needs have been neglected. National policy papers should be prepared on both issues to forge consensus and guide implementation. Earlier government environmental assessments are needed to influence project design, and environmental monitoring needs to be enhanced. Successful pilot resettlement and rehabilitation schemes now need to be generalized across all projects, the key being detailed preparation, full funding as part of project costs, and early establishment of instituional capability. vii. The technical skills for quality maintenance and construction are generally well known. fhe deleterious element is inadequate control of quality due to lax supervision and monitoring exacerbated by rent-seeking influences. Much tighter management of maintenance and construction is required but resolution goes beyond the orbit of the supervising engineer and will require commitment and backing from senior state authorities.  iv xviii. Non-GoveMment/Government Roles and Capacity. The greatest challenge is in the management domain, the source of most of the sector's problems due to weak capacity, increasingly overwhelmed by the growing complexity of the sector in the face of limited institutional adaptation and encroachment of rent-seeking and often disruptive polit.%al pressures. Stae-level improvement will have greatest influence, particulariy within the irrgation department, the principal implementing agency. Numerous measures are available to improve irigation department management: creation of a corporate management curture and integration of non- engineering skills; organizational restructuring along functional lines (e.g., maintenance, water management, monitoring, project preparation and design); aggressive training and use of consultants to upgrade expertise; and reporting procedures to increase public accountability. The goal should be to shift inigation department managements from a naffow engineering focus to a managerial one that permeates all functions and levels. Clarification of roles and better coordination with other deparments involved with irrigation are also required. xix. Parallel steps are needed toward progressive divestiture, where feasible, of functions and implementation to the non-government sector. Short-tern oppoitunities are available. The resources of universities, research centers, NGOs and consultants can be used to enhance implementation capacity, research and training. Maintenance can be contracted to the private sector rather than implemented by government as commonly done. Government functions at the local level (junior water management staff, water charge collectors) can be fulfilled by water-user associations or panchayats, and public functions eliminated. xx. For the longer term, actions should be charted now to create autonomous commands. Commands will only achieve financial discipline over maintenance and operating costs and become responsive to user needs if they are financially independent and managed by users. A major thust is required to help establish water-user associations, to move fanner responsibility progressively to higher levels of command management, to create independent entities (e.g., registered societies), and eventually to hand over functional control to farmers, with water charges retained by the command. Suitable commands for initiating this approach are to be found ir. most states. xxi. Immediate Action. Over the next 12 months, the states should execute the following activities on the path to a clearer investment focns, fiscal responsibility and improved productivity and sustainability of the sector: o Re-examine expenditure focus and develop new prorties based on economic viability and impact o Prepare and implement an interim action plan to halt staff growth, provide adequate maintenance funding, curb leakages and introduce cost accounting; o Commence annual reporting of financial expenditure and cost accounting, a maintenance plan, budget and review, and a cost recovery status report; o Prepare and initiate a water management improvement program; o Begin restructurng of irrigation departments, injecting non-engineering skills, establishing an MIS function, upgrading project preparation capabilities and funding training and consultancies; o Identify areas for private sector involvement; o Identify specific commands where management can devolve to autonomous units; and o Initiate an extension program to assist farmers to form water-user associations, paroticipate in scheme management, and maintain and operate micronetworls.  v For its part, the central goverr nent should: o Strengthen its review function of irrigation planning and finance by the states, insisting on good maintenance, finpiicial discipline and expenditure priorities; o Promote the national water policy and creation of river basin planning commiissions; o Prepare national policies on environment and resetlemelht and rehabilitation; o Strengthen inter-ministerial/agency cooperation and staff capabilities focussing on water policy coordination, overview of state perfcrmance, strategic analysis, research networking and project evaluation; and o Identify programs for central financial support to state initiatives in priority functional areas (e.g., water management). xxii. The need is clear. Improved irrigation performance will depend in the short term on the degree to which managerial discipline can be brought to bear on the sector at all levels: inigati'rn departments, state governments and at the center. Further improvement will depend on more fundamental charne, recognizing that government cannot do everything, that the initiative and capital of the non-government sector should be tapped, and that management of completed schemes should become the administrative and financial responsibility of users. Some steps that will be necessary to lead in the prescribed direction are bold, requiring restructuring, reform, and progressive devolution; others are pedestrian, requiring more diligence and tenacity in current roles and functions. In both cases, the outcome should be improved productivity in irrigation, a critical need in India for the remainder of this century and beyond. xxiii. Postscr. The past two years have involved substantial reflection within India, at both central and state government levels, on the future needs of Indian irrigation. This Irrigation Sec Review, undertaken in collaboration with the Indian govemment, and govenmment's ongoing preparatory work for its forthcoming Eighth Plan (1992-97) have been focal points in these reflections. The September-October 1991 discussions between central and state governments and the Bank of the draft ("green %,over") Irrigation Sector Review report provided an additional opportunity to further establish consensus on sectoral needs. This final report has been adjusted to reflect the substantial consensus now present in india. While some differences in detail or timing reg3rding the recommendations present in the report may remain, there is substantial unanimity of thinling on the report's diagnostic and recommendations. Initiatives in the prescribed direction are already underway at both central and state government levels, and have as far as possible been referred to in the final :eport. xxiv. Specifi. actions needed will vary by state, reflecting regional variations in irrigation development opportunities available, agroecology, and in particular, the status of current irrigation sector performance. The magnitude of adjustments needed in most states seldom lcnd themselves to instant full remediation; rather, improved irrigation performance will require a sustained process of change over time. The turn-around in sector performance that this drive can achieve will itself demand strong political will and back-up commitment from central and state govemments, and support towards such efforts from the donor community.  CHAPTER I. THF CHALLENGE AND STRATEGIC CHOICES A. Critical Issues 1.1 Agricultural performance is fundamental to India's future economic and social development. Agriculture contributes 30% of GNP, 60% of employment, and is the primary source of livelihood in rural areas which account for 75% of India's population and 80% of its poor. With all arable land under cultivation, increased agricultural output will depend on raising crop yields, increasing cropping intensity, and diversification to higher-value crops. The performance of irrigated agriculture, which contributes 55% of agricultural output, will be the most important influence on these objectives. Addressing the irrigation sector's current performance problems will thus be a central element of future strategy for agricultural development. 1.2 'WVith the largest irrigated area in the world, Indian irrigation has much to be proud of. Its development, the product of major efforts and achievements by India's irrigation engineers and of greatly expanded government i.ivestment since Independence, has been the principal force behind agricultural growth. Its role will be indispensable to future growth. Nonetheless, due to rapid expansion of irrigation with its emphasis on new construction, the performance of irrigation and the sector's broader management needs have been neglected. The development impact of irrigation is much less than its potential and deficiencies in implementation have accumulated over time. The sector is now in crisis. Four issues are of particular concern: productivity, sustainability, investrrent focus and financial discipline, and sector management. 1.3 The productivity of irgation is low and well below potential. On the majority of surface irrigation schemes, suboptimal distribution of water (timing, reliability and spatial distribution) results in low yields and cropping intensity and reduced opportunities for diversifying agriculture. Deficiencies in agricultural extension, limited research on irrigation technology, and insufficient piloting of innovations compound the situation. Poor design and project preparation of many surface irrigation schemes and deteriorating infrastructure are also having significant deleterious effect. 1.4 The sustainabilitv of India's vast irrigation investment is in serious question, mainly due to a marked decline in maintenance of infrastructure and a similar decline in the quality of construcdon over the past two decades. Rehabilitation requirements are already large and threaten to dominate expenditure unless maintenance and construction are improved. Additionally, environmental problems are mounting, requiring substantial attention to them, drainage, and resettlement of displaced communities. A void in water resource planning and coordination of usage urgently needs to be filled. 1.5 The investment focus and financial dcipline of irrigation require rethinking and reform. Substantial reallocation of government expenditures is required to support investment for improving productivity, adequate maintenance, completion of ongoing projects, drainage, and micronetwork and groundwater investment by farmers. Sihzap financial deterioration of the sector poses a serious threat. Recurrent expenditures have ballooned in the past decade, with most of the increase going to unproductive staff growth and wastage due to lax control of expenditures and works. 7st recovery has dwindled and irrigation-related subsidies now loom large in state government budgets. These impacts are crowding out worthwhile investment. 1.6 Ineffective sect ilianagrmenl underlies most of these deficiencies. Sector institutions, notably state irrigation departments, have undergone litde change over many decades, while sectoral priorities have changed and execution has become more complex. The departments' structures, functions and approach are predominantly construction-engineering; this orientation is inadequate in face of the need for a broader management culture and sharper diagnostics and response to the sector's weaknesses and opportunities. The roles of other state institutions involved with irrigation need better definition and coordination among them. The vitality and  2 capital of non-government sources are not being captured, and significant opportunity for greater involvement of farmers and the private sector has been virtually ignored. 1.7 If these issues are not resolved, productivity will iemain low; infrastructure will progressively deteriorate, leading to declining irrigated area or insupportable rehabilitation needs; financial imbalances will continue to widen; and most seriously, there will be little scope for maintaining the agricultural growth rate. 1.8 It is clearly within the capabilities of the sector's managers and its engineering cadre to tackle the problems, as evidenced by many examples of excellent performance in Indian irrigation, even though limited in number and regionally concentrated. Major improvement is required and no state is entirely free of problems discussed in this review. The proposed strategy calls for a substantial reform of irrigation sector management, focussing on critical performance problems. Reform will only be achievable with strong consensus and commitment from senior political and government establishments at both state and central levels. B. Development of Irrigation 1.9 Investnent in irrigation has been a development priority in India.l By 1950, 21 million hectares (ha) of land were under irrigation, mostly built during the previous 75 years, but some works dating to before the 10th century. Successive five-year Plans have emphasized irrigation investment, with expansion of net irrigated area averaging 0.6 million ha per annum. Net irrigated area had doubled to 42 million ha by 1984-85. With part of this area double-cropped, gross irrigated area was about 54 million ha or nearly one third of India's gross cropped area.2 Table 1.1 Development of Irrigation: 1950 - 1985 (Net Irrigated Area in Million Hectares) 1950-51 1960-61 1970-71 1980-81 1984-85 Canal Irrigation 8.3 10.4 12.8 15.3 15.9 Tanks 3.6 4.6 4.1 3.2 3.3 Tubewells n/a 0.1 4.5 9.5 11.3 Other Wells 6.0 7.2 7.4 8.2 8.7 Other Sources 3.0 2.4 2.3 2.6 2.6 Tot. Net hbrig. Area 20.9 24.7 31.1 38.8 41.8 (Growth Over Previous Period; % p.a.) (N.A.) (1.7) (2.3) (2.3) (1.9) Source: Directorate of Economics and Statistics, Ministry of Agriculture. 1.10 Rapid growth of irrigation since 1950 has involved both construction of new surface irrigation schenms and increase in groundwater irigation (Table 1.1). Between 1950-51 and 1984- 85, public investment in canal irrigation doubled the area under surface irrigation from 8.3 million ha to 15.9 million ha (about 220,000 ha or 1.9% per annum). Since the early 1960s, private tubewell irrigation has grown spectacularly, from almost nil to 11.3 million ha by 1984-85 (7% per I Refer Volume II, Chapter I for more detailed background and analysis. 2 Throughout this repor, the terms "net" and "gross," utilize Indian rather than international nomenclatue. "Net irrigated area" refers to land area. "Gross irrigated area" refers to cropped area Thus, where 1.5 irrigated crops are grown per annum on a 1,000 ha irrigated parcel of land, net irrigated area is 1,000 ha and gross irrigated area is 1,500 ha.  3 annum growth between 1970-71 and 1984-85). More modest growth in the number of dugwells brought the net area irrigated by well irrigation to 20 million ha by 1984-85. 1.11 Irrigation has played a pivotal role in India's "Green Revolution." The spread of high- yielding wheat and rice varieties in the 1960s and associated use of fertilizer was made possible by increased volume and reliability of water supplies from irrigation. Irrigation also increased cropping intensity due to possibilities of growing crops outside the monsoon season. 3 The Green Revolution sustained agricultural growth at an average rate of 2.5%. While not spectacular, it was achieved in the face of limited productivity growth in rainfed agriculture. Agricultural growth was, however, heavily concentrated in the irrigated northwest and some irrigated areas in peninsular India where the water regime also penritted intensification. 1.12 Agroclimate. Irrigation in India must cater to widely diverse climates and socioeconomic conditions. Rainfall is seasonally concentrated, with most areas receiving 70-95% of mean annual rainfall from June to September. This concentration, coupled with dry periods during the monsoon, makes irrigation important for intensive agriculture in most regions. In the northwest (Punjab, Haryana and parts of adjacent states) the climate is dry with rainfall of less than 600 mm, but dry season river flows are supported by snow melt from the Himalayas. In eastern and central India (east Uttar Pradesh, Bihar, West Bengal, Orissa, Assam, and much of Madhya Pradesh), rainfall ranges between 900-1500 mm and even higher in some localities, with monsoon flooding a common problem. Most of peninsular India (Andhra Pradesh, Karnataka, Tamil Nadu, most of Maharashtra and southern Madhya Pradesh) has seasonally concentrated rainfall of 600-800 mm. Specific localities are affected by proximrity to mountains and oceans; in Kerala, rainfall exceeds 1500 mm. High rainfall also occurs in the Himalayan foothills. Arid conditions prevail in westem states (Rajasthan and Gujarat) and, to a lesser extent, in some rainshadow areas in peninsular India. 1.13 Monsoon intensity, duration and reliability influence the nature of irrigation reqdired. In wetter regions, occasional but well-timed supplemental irrigation during dry spells of the "kharif" (monsoon) season is all that is required, while more regular irrigation is needed during "rabi" and summer seasons. In dry regions, irrigation is the main source of water determining crop growth, even in the kharif season. Other relevant physical factors are soils, topography and temperature. The degree to which groundwater is available is a particularly important factor, providing the option of well irrigation either by itself or in conjunction with surface irrigation. 1.14 Regional Socioeconomics. Substantial differences in socioeconomic conditions and traditions also affect irrigation's performance. In eastern India, famns are small and fragmented and often under sharecropping or other tenancy systems. Yields are generally low and extreme poverty is widespread. Infrastructure, such as roads, electrification and availability of govermment agricultural support services, inputs and markets, is also poor. More favorable socioeconomic conditions prevail in northwestern states such as Punjab where population pressure is lower, farm holdings are larger and generally owner-occupied and unfragmented, and infrastructure and institutions are better developed. 1.15 Diversity between State. Agroclimatic, socioeconomic and historical development differences between states have all fed into substantial variation between states in the dimensions and nature of irrigation, the performance of irrigation and its impact on agricultural production. India must be recognised as a large and diverse country and this is reflected in the diversity of experience and development needs of its irrigation sector, with irrigation experience ranging from exemplary to poor. Future needs for improvement and development of irrigation reflect this diversity; between regions, states and even within states. Such differences are highlighted as 3 Irigation development, while crucial, was only one of the factors affecting agricultural development Other factors include availability of suitable improved varieties, fertilizer and other inputs, and support services such as extension, research and credit. Rual infrastructure such as roads, availability of markets, electrification, schooling and health services have also played an important role (Binswanger, 1989).  4 feasible in this report, with greater regional and state-level commentary in Volume II. The report's recommendations, though most have application throughout India, will also require adaptation to the specific circumstances and needs in each state. C. Impact of Irrigation 1.16 The development of irrigation has fundamentally influenced agricultural productivity, incomes, employment, and regional development. However, impact has varied consiterably, depending on type of irrigation, agroclimate and, above all, on the technical quality and management of irrigation. Principal impacts (Vol. 11, paras 1.6-1.39) are summarized below. 1.17 Croping Intensity and Exoansion of Cropped Area. While constraints on arable land prohibited expansion of net area under cultivation in the past two decades, gross cropped area expanded from 166 million ha in 1970 to 177 million ha in 1985, with an estimated 60% of this expansion due to irrigation. Still, average irrigation intensity remains low (1.29 irrigated crops per annum in 1984), due to the prevalence of run-of-the-river irrigation systems dependent on monsoon rains and primarily providing water for protecting the kharif crop. The main impact has been due to growth of groundwater irrigation, which markedly increased cropping intensity in regions such as the northwest. 1. 1 8 Crop Yields. Average yields under irrigation are more than twice as high as yields under rainfed conditions (for cereals, 2t/ha, compared to 0.8t/ha rainfed), but low relative to yields under irrigation in other countries. This is mainly due to poor water management on the majority of the surface command area. Good irrigated yields are achieved (averaging 3t/ha/crop for cereals in the northwest and much higher for individual farmers and in some commands with ample water in parts of peninsular India), but yields are low in eastern and central India and in many commands in peninsular India. Farmers with access to water through private tubewells or dugwells generally have good to excellent yields. 1.19 Output Stabilization. Irrigation has helped reduce inter-annual fluctuations in agricultural output and India's susceptibility to droughts. Enhanced by improved transport linkages and Government's buffer stock system for food staples, the serious famines of the past are now less likely.4 1.20 Crop Diversification. Irrigation has only been selectively responsive to the needs of diversified agriculture. Combined with high-yielding varieties and puice supports, it principally affected an increase in cultivation of wheat and rice, together accounting for two thirds of expansion in gross irrigated area since the 1960s. This overshadows the doubling of non-cereal area under inigation, particularly of oilseeds, sugar, cotton, fruits, vegetables and dairying. Diversification to these crops, especially to fruits and vegetables, has been mainly facilitated by development of groundwater irrigation: private tubewells and dugwells, and in recent years, rapid growth of drip irrigation in peninsular India. 1.21 Farm Incomes and Employment. On average, farm incomes have increased 80-100% as a result of converting from rainfed to irrigated farming. Incremental labor days used per hectare 4 Commentators in the early 1980s (Mehra, 1981 and Hazell, 1982) noted a possible adverse impact of irrigation on output stability, but later analysis using more complete data (e.g., Dhawan, 1988) indicates an aggregate stabilizing impact, especially in regions of lower rainfall and where irrigation and drainage have been successful. At local levels, tiiis may not be the case, in particular where farmers take up more intensive (more water- regime susceptible) agrculture and where irrigation unreliability and flood prospects, for instance in eastern India. are still presenL  5 average 60-80%. On larger farms, some of this represents incremental employment of non-family labor (usually from landless households or families with marginal farms). On smaller farms, incremental labor requirements tend to be handled by the family.5 1.22 Secondary Income and Employment Effects. Through its influence on agricultural incomes, irrigation development has a multiplier effect on non-farm incomes. Nonagricultural activities such as rural retailing, artisanal and industrial activities, processing and services grow as agricultural output and incomes rise. A 100-rupee increase in irrigated agricultural output stimulates 105 rupees of additional output in manufacturing and 114 rupees of additional tertiary output, a total nonfarm output multiplier of 2.19 (Hazell and Haggblade, 1990)6. Expansion in agricultural output as a result of irrigation has also helped keep food prices down. Between 1970 and 1986, food grain prices in India fell by about 20% relative to the price index for all commodities. This has had significant impact on the real incomes of the urban poor and landless rural households for whom food is a large component in consumption. 1.23 Poverty Alleviation. Analysis of district-level data across India shows a strong relationship between agricultural productivity and the proportion of land area irrigated and associated usage of farm inputs (Bhalla and Tyagi, 1989). Similarly, the expected inverse relationship between the incidence of poverty and the extent of irrigation development has been demonstrated in district- level comparisons (Rao, 1988). For districts where less than 10% of gross cropped area was irrigated, 69% of the population had incomes below the poverty line, while in districts where irigation covered more than 50% of crop area, poverty incidence was only 26%. 1.24 Regional Development. Regional impact has been dependent on availability of water resources and quality and extent of irrigation development. Past investment has been regionally concentrated, reflecting both water endowment and past policy emphasis on irrigation development in northern states. Since Independence, efforts have been made to broaden regional coverage of irrigation, especially in peninsular India, but given the distribution of water resources, such efforts will have limitations. Improving rainfed agriculture will, thus, be a continuing priority, as irrigation potential is inherently imbalanced by region. 1.25 Water resource endowment and investment have favored the northwest where irrigation has transformed agricultural productivity and general wealth. More than half of net cropped area is irrigated (82% in Punjab and 62% in Haryana). Between 1962 and 1982, the Punjab's net cropped area under irrigation increased from 53% to 82%, enabling rapid uptake of high-yielding varieties and fertilizer, and making the northwest the main beneficiary of India's Green Revolution. Other agricultural developments, including use of farm machinery, followed. While irrigation was not the only infrastructural influence, development would have been impossible without it. Under rainfed conditions (less than 600 mm), the land is only suitable for extensive pasture or drought- tolerant crops, such as sorghum and millet, with probable annual yields of less than one ton per hectare. Irrigated yields in Punjab average 3 tons per crop (rice and wheat) or 6 tons per annum. As a result, agricultural growth averaged 4.5% per annum during 1963-82, and Punjab's agricultural productivity is now more than twice that of India as a whole. In turn, agricultural growth has also stimulated the northwest's non-farm economy. 1.26 By contrast, agricultural growth in eastern and central India (48% of India's net cultivated area and one third of net irrigated) is less than 1% per annum. As share of net cropped area, irrigation currently covers 12% in Madhya Pradesh, 20% in Orissa, and about 35% in Bihar and West Bengal. Yields are low, despite substantial rainfall. The major factor is that poor-quality surface irrigation schemes have not proven responsive to seasonal crop needs. For the typical rice- 5 The average impacts on employment and incomes noted above vary widely, depending on the effectiveness of irrigation (Vol. 11, paras 1.22-1.26). 6 The multiplier is higher in more developed states because of higher consumption linkages and input intensity.  6 wheat rotation, water is required before the monsoon, but most surface schemes can only deliver water after the monsoon is well underway, and water distribution is unreliable, not responsive to dry periods, and unevenly distributed between head- and tail-ends of commands.7 A minority of fanners have bypassed these constraints with tubewell investments and obtain yields comparable to those in the northwest, but groundwater irrigation, while expanding rapidly in Uttar Pradesh and parts of Bihar, is still limited in most of eastern and central India.8 1.27 In peninsular India, rainfed agriculture prevails and is likely to remain dominant. In the west (Gujarat, Rajasthan, Maharashtra), less 'han 20% of arable land is irrigated. In the sqlut (Tamil Nadu, Andhra Pradesh, Kamataka and Kerala), net area irrigated is 48%, 32%, 20%, and 12%. respectively. Irrigation's impact is variable, ranging from excellent yields in some commands with ample water to the more prevalent situation of mediocre to poor yields due to weak water management. A common situation is unequal distribution of water, head-enders getting plenty to the detriment of tail-enders getting little or none. D. Potential for Irrigation Development 1.28 Irrigation can have major impact on agricultural productivity, growth and on general development and poverty alleviation, as achieved in northwest India, in selected ccimnands elsewhere, and with groundwater irrigation. Impact could be higher in all states and much higher in most states, particularly in eastern and central India and on many commands in peninsular India. Even successful irrigation in India has scope for improved productivity. Regional development concerns, especially for poverty areas, must also be considered. Irigation and Drainage 1.29 Irrigation's development potential cannot be stated precisely, due to acknowledged limitations in available data, lack of detailed river basin plans, separate estimation of surface and groundwater resources, uncertain assumptions regarding competing demands for water, and other factors.9 The Central Water Commission's (CWC) 1985 assessment is the most recent official estimate (see Table 1.2). As recognized by CWC and discussed elsewhere in this review (Vol. II, Chapter I1 and Barber, Patel e a!., 1990), conservative interpretation of the official assessment is in order. 1.30 Official estimates put irrigation potential at 45.6 million hectares (113.5 million ha of ultimate potential minus 67.9 million ha of potential created). Of this, 28 million ha is estimated to be suitable for major and medium surface irrigation development and .7.6 million ha for groundwater irrigation. An additional source of expansion includes closing the gap between potential created and actual potential, estimated by government at over 7 million ha. While 7 Timing is very critical. To obtain a good rice crop, water is needed before the monsoon to allow tillering (for good crop density) and fertilizer application before waterlogging. Earlier rice planting also enables earlier harvesting and then planting of wheat for growth during the cool months, critical for good wheat yields (India Agricultural Technology Review, 1989, India Agriculture Operations Division, World Bank). 8 The constraints in eastern and central India are complex, involving, in addition to poorly perfonning irigation and drainage, poor infrstructure, socioeconomic factors such as land tenure and caste structure, and weak institutions. Inadequate control of water is, however, a basic factor, also affecting the possibilities for using high- yielding varieties and fertilizer (Pray, 1991). Other commentaries are in: Agricultural Productivity in Eastern India, RBI, 1984; and Report of Study Group on Agricultural Strategies for Eastern Region of India, Planning Commission, 1985. 9 For example, dependency on assunptions regarding capturable surface run-off, the rate of groundwater recharge, irrigation efficiency, and the intensity of water usage on irrigated lands.  7 dimensions of the gap are open to question,l1 there is large scope for expansion through improving water distribution on existing schemes, construction of micronetworks and completion of ongoing projects. A second source of expansion, not included in 'Table 1.2, is the transfer of water from river basins with suiplus water to water-short basins. Aggregate potential from such schemes has been broadly assessed by CWC and NWDA to be up to 25 million ha. II Such transfers would be expensive and require water-sharing cooperation between states. If technically and economically viable, transfer schemes to water-short southeastern and western basins may be necessary over the long term. Table 1.2 Government Estimates of Status of Indian Irrigation Development in 1985 Major & Minor Medium Scheme Schemes Total ----- million ha ----------- Ultimate Potential (1981 estimate) 58.5 55.0 113.5 Potential Created (at 1985) 30.5 37.4 67.9 Potential Utilized (at 1985) 25.3 35.2 60.5 Balance (Ultimate minus Created) 28.0 17.6 45.6 Note: Additional potential from river-basin transfers (para 1.30) are not included in ultimate potential. Source: CWC 1.31 There is a clear need for stronger hydrological data and revised assessment of development prospects based on water's natural hydrological unit, the river basin. Until then, assessment of potential can only be indicative. Realizing this potential will be more technically difficult and costly than in the past and most opportunities will depend on creating additional storage capacity. Potential is also regionally concentrated and must cater at local levels to competing demands for water use. Nevertheless, within these uncertainties and constraints, prospects for expansion are present, as discussed below:12 1.32 Groundwater Irrigation. Prospects for further groundwater irrigation, while less than official estimates of 17.6 million ha, are still significant. In many of India's river basins, there is substantial scope for continued rapid expansion of groundwater irrigation over the next 10 and perhaps 20 years. The greatest potential is in eastern India. Moderate but localized potential also exists in western and southern states. However, in parts of the northwest and west, exploitation has reached its limit. In dry areas with limited natural recharge, groundwater irrigation prospects are largely of a conjunctive nature, dependent on recharge due to availability or expansion of surface irrigation. Its future is thus closely interlinked with development of surface irrigation. 1.33 Suirface ligation. Possibilities for new run-of-the-river schemes are virtually exhausted and investmen' in new surface irrigation will be almost entirely dependent on creation of additional water storage capacity. Capturing surface runoff is particularly important in India, with most 10 Potential created may be overestimated because inigation efficiency estimated at design is often too high, further compounded in practice by farmers using cropping patterns requiring more water than originally assumed. 1 1 The National Water Development Agency (NWDA), established in 1982, is charged with "n4ikdng assessmnents of water resource development prospects from inter-basin transfers. 12 For more detailed discussion and assessment, refer to the approximate river basin and state estimates in Vol. II, Chapter I and its supporting annex and to "Assessment of Indian Water Resources and Irrigation Development Potential," W. Barber, C.C. Patel gal. (Irrigation Sector Review Background Paper, 1990).  8 rnvers carrying about 80% of their flow during the monsoon months. Estimates of surface irrigation potential (28 million ha) assume high uti!ization of natural runoff, but this will be difficult to achieve in practice as good dam sites for creating additional storage are limited in most basins. Even so, at the current expansion rate (about 220,000 ha net per annum) or at a faster development rate, sufficient opportunities are available in India for expansion of surface irrigation through several decades 1.34 Investments in storage reservoirs can be expected to have higher development costs and marginal rates of return (para 1.50). Further, the issues of finding satisfactory solutions to environmental problems and resettlement and rehabilitation of populations displaced by reservoir developments will be more prominent. Additionally, many future options would require interstate agreements on water-sharing. Planning and administration of projects that cross state boundaries are already a problem. Despite these issues, investment for the future in water storage is essential now in certain areas, and will be the basis for longer-tern water development throughout India. 1.35 Drainage. Investment in drainage has been widely neglected, and where such investment has been made, poor maintenance has caused many drainage systems to become silted up. A large part of both rainfed and irrigated eastem and central India, most of the eastern deltas, and localized bottomlands of irrigated commands elsewhere in India are subject to waterlogging and flooding during the monsoon. In addition to crop failure, the more widespread problem is that excessive water imposes limitations on varietal choice, fertilizer application, and cropping patterns, all affecting yield. In the eastern Gangetic plain, investment in surface drainage is likely to have larger productive impact, and at lower cost, than investment in surface inigation.13 1.36 Additionally, parts of northwest India and Uttar Pradesh (250,000-500,000 ha) are affected by a build-up of saline groundwater; in extreme conditions land has gone out of production. In these areas, more expensive groundwater drainage is required if agriculture is to remain sustainable. Regional Development 1.37 Norhwest Only limited extension of irrigated area is possible in the northwestern states of Punjab and Haryana and parts of adjacent states where potential is fully utilized, especially in Punjab. The Punjab Lrrigation and Drainage project (Cr. 2076/Ln. 3144) will complete marginal extensions for surface schemes, improve efficiency of surface water delivery, and improve groundwater drainage. Some additional surface water irrigation is feasible for Haryana if storage is constructed on the Yamuna in the Uttar Pradesh hills. Modest additional groundwater development is possible in localized areas of Haryana, while in almost all of Punjab, groundwater is fully developed. However, major additional investment is required in surface and groundwater drainage in southwest Punjab, Haryana and in adjoining areas in other states to combat increasing waterlogging and soil salinization. 1.38 East and Ce_n_tr. The largest unexploited irrigation potential is in eastern and central India. Uttar Pradesh, Bihar, West Bengal, Orissa and Madhya Pradesh together are estimated to have 56% (up v) 25 million ha) of the country's untapped potential. 14 The most attractive opportunity is for groundwater development (40% of east and central India's irrigation potential). However, while good groundwater prospects exist in all these states, development needs to be confined to areas where exploitation would not adversely affect dry-season Ganga river flows, i.e., where the water table is high and subject to monsoon waterlogging. Major opportunities are available for improving water management on surface irrigation schemes, investing in drainage and flood protection to reduce risks and improve kharif yields in areas with excessive wet season water, and tackling the problem of saline and sodic soils in Uttar Pradesh. For many years to come, 13 "Land Reclamation and Drainage," L. K. Smedema, (Irrigaton Sector Review Background Paper, 1990). 14 As at 1985, based on GOI's assessment.  9 expansion of surface irrigation can focus exclusively on completion of the substantial backlog of ongoing projects, especially in Madhya Pradesh and Orissa. 1.39 South. In 1985, only 6.5 million ha of irrigation development potential was estimated (14% of India's remaining potential) for southern India, and this is likely to be an overestimate. Over half is in Andhra Pradesh. Tamil Nadu has fully developed its water resources, while the demands of conurbations such as Madras are increasing. However, some scope exists for localized groundwater irrigation and better economies in water usage on existing surface commands.15 In Andhra Pradesh and Karnataka, completion of ongoing surface schemes and improving water management of existing scheme are priorities. Some it-ientiai exists for creating additional storage on the Godavari in Andhra Pradesh, and there is moderate scope in both Andhra Pradesh and Karnataka for additional groundwater development. Investment in drainage may be attractive in irrigated areas of the southern states, especially on the waterlogged and flood-prone delta schemes. Longer-term expansion of irrigation in southern India will depend on the technical and economic feasibility of water transfers from surplus basins further north and across the Western Ghats from Kerala. 1.40 West. The states of Rajasthan, Gujarat and Maharashtra have moderate remaining potential for irrigation development, some 6.8 million ha, based on the 1985 estimate. Both existing and potential irrigation is highly localized, leaving large areas of these states still dependent on rainfed farming. Groundwater potential exists in all three states, especially Maharashtra, but may be only 25% of the region's total remaining potential and in many localities is dependent on conjunctive use with surface irrigation. The bulk of surface irrigation potential is represented by incomplete, ongoing projects and the Narrnada development program, particularly significant to the region's developmnent prospects. The Narmada's Sardar Sarovar projects (Cr. 1552/Ln. 2497 and Cr. 1553) will service rapidly growing municipal and industrial needs, furnish hydroelectric power to the regional grid and provide irrigation for a 1.8 million ha command area in arid Gujarat. E. Economic Viaility of irrigation and Dainage Investment Past Perfomance and Viablity 1.41 The economic viability of irrigation investments in India has been critically affected by implementation, which has been weak in most cases, particularly so on government-financed projects. Economic rates of return are not calculated by government, so this assessment must be based on comparison of performance relative to World Bank-assisted projects; despite many deficiencies, Bank projects have tended to fare better. Notwithstanding efforts to undertake more thorough project preparation, basic weaknesses in data, project design, detailed engineering and planning have resulted i. inappropriate project features. Adjustments to rectify deficiencies during implementation have not been easy to make and have contributed to implementation delays. Project appraisal has often been conducted before all design and institutional features have been finalized or resolution of land allocation, resettlement or environmental issues, also contributing to delays and poor implementation. Additionally, during implementation, inadequate government supervision of contractors has caused delays, substandard works requiring reconstruction, and cost overruns. Poor coordination across government departments has also impeded execution and reduced quality. And, most importantly, once constructed, most irrigation schemes have been poorly operated and maintained, other services (extension and research) have been inadequate, and yield potential seldom realized. 1.42 As a result, re-estimated economic rates of return (ERRs) for World Bank-assisted projects at completion have mostly been poor (Volume II, Chapter III). Of nine projects subject to project 15 This is, however, limited as many schemes (including the tank irrigation systems) are in cascade; losses on one scheme are utilized on another downsteam scheme.  10 completion reports (PCRs) in 1989, all with satisfi>.tory ERRs estimated at appraisal, only two had re-estimated ERRs above the estimated opportunity cost of capital in India of 12%. Four had ERRs of 5% or less. Although a decline in world rice prices in the mid-1980s had some impact, poor implementation was the primary influence. In seven projects, capital investment costs were substantially higher than anticipated, with cost increases of 49-147%. Seven projects also experienced implementation delays, substantial in several cases. Other important factors, most still not fully captured in the economic analysis of PCRs, were reduction; , command areas expected to be irrigated, shortfalls in yields, and downward adjustments in cropping patterns assumed at appraisal. Similar results were noted by Daines and Pawar (1987) in economic analysis of government-funded projects. More often, performance problems have been greater on government- funded projects, due to hastier project preparation, less attention by state and central governments during implementation, and funding constraints that slow implementation. 1.43 Despite substantially better ERRs in PCRs on earlier World Bank projects, there is little to indicate that recent results are exceptions. Based on field observations and anal-sis of the methodology used for calculating ERRs in earlier PCRs, the most reasonable concluF ion is that project performance and economic viability have been poor all along for most projeccs. 16 Of 13 projects completed between 1970 and 1984, two had ERRs of less than 12% and half had ERRs greater than 20%. However, half overran cost estimates at appraisal by 50% or more. Moreover, PCRs did not take into account that yields and actual irrigated area were likely to be smaller than assumed at appraisal, even when field observations might have indicated that this was likely to be the case. Rising rice prices also improved PCR rate of return estimates during that period. Returns by Type of lITigation Investment 1.44 What kind of investments in irrigation and drainage have maximum viability as future investments? The observations below, based on ex-post and ex-ante ERR indicators, provide some guidance.17 1.45 Groundwater Irrigation. Groundwater irrigation with small private shallow tubewells and to a lesser extent with dugwells, each serving about 3-4 ha, have been estimated to have ERRs in excess of 30% (Daines and Pawar, 1987). Even with adjustment of electric power costs, which may be underassessed here, economic returns are high. Private tubewells enable farmers to control water so that irrigation can be tailored to crop needs. Financial returns to farmers from groundwater development are even greater than the economic returns, as investment costs are partly subsidized and electrically operated pumps highly subsidized. Although diesel fuel is unsubsidized, diesel-operated tubewells also have high financial returns, as benefits from high yields substantially outweigh investment and operating costs.18 1.46 Public tubewells (deep tubewells constructed, maintained and operated by government and designed to serve farmers in an 8-100 ha block) are much less viable. Appraisal of World Bank- financed projects in Uttar Pradesh, Bihar and West Bengal put estimated ERRs at 30-40%, but assumptions about command area served, yields, and cropping intensity have proved to be too optimistic. Field observations indicate modest impact on average productivity and difficulty in sustaining these investments, which can be expected to be reflected in ex-post ERRs. The ongoing Indo-Dutch tubewells project in eastern Uttar Pradesh, modelled after the World Bank project (Cr. 16 Re-evaluation of several earlier PCRs is currently being undertaken by the Operations Evaluation Department, World Bank. 17 Wherever possible, ex-post ERRs have been used. Where ex-ante ERRs are compared against each other, the assumption is that they provide some indicator of comparative economic viability in a situation where performance is reasonable. Refer Vol 11, Chapter III for more detailed analysis. 18 Diesel prices, although regulated, are adjusted periodically to approximate international prices at the official exchange Mte. A further upward adjustment to reflect exchange rate distortions would be desirable.  I1 1332), has a re-estimated (1990) ERR of 9%, using more realistic assumptions based on current experience. Recent World Bank analysis concludes that fostering private groundwater irrigation with shallow tubewelis is the more attractive development option.19 1.47 Existinmg Suface irrig Available data support the economic viability of expenditures to improve performance of existing surface irrigation systenms. Substantial sunk costs have. already been incurred, and the incremental cost of institutional improvements and infrastructure are relatively low, compared to substantial potential gains in productivity and efficiency. Improvement options include: better operations management; more intensive modernization of irrigation schemes; development of micronetworks (watercourses, field channels and field drains); and upgraded agricultural extension and adaptive research (on crop varieties, cropping systems and irrigation technology). 1.48 For example, the National Water Management Project (Cr. 1770, para 4.26) aims at improving efficiency of water distribution, mainly by upgrading planning and management and some water control structures. At appraisal, ERR estimates were over 30%, and encouraging initial results are being encountered in some states. Projects that involve more intensive investment to upgrade and modernize the irrigation system, ranging from water control systems to canal lining, also have generally higher ERRs at appraisal than those for new surface irrigation. Actual returns have varied, depending on implementation: the Punjab Irrigation Project (Cr. 889) yielded an ERR at completion of 36%, while the second Chambal project in Madhya Pradesh (Cr. 1288) achieved 10%. The viability of canal lining depends on physical features and the quality of lining. Where seepage beneficially recharges groundwater without waterlogging problems and such water can be retrieved by tubewells, lining is often not justified. In all cases, quality must be good (and often has not been) if lining is to be beneficial. Less analysis is available on the impact of micronetwork development, but high economic viability can be expected. Irrigation cannot fulfill its function until the tertiary system is developed and well functioning, still a major weakness on most commands. Similarly, provision of water must be accompanied by good agricult'lral extension, another common weakness, and applied irrigation technology and agricultural researn. are fundamental to improving irrigation and agricultural yields. 1.49 Drainage. Investment in drainage infrastructure can be expected to have good economic viability. In the eastern Gangetic plain (both rainfed and irrigated areas) and in many deltas, as well as localized bottomlands in suiface commands elsewhere, surface drainage can substantially reduce waterlogging problems and the duration of temporary flooding. Good impact can be anticipated on crop intensification a-id diversification opportunities available to farmers. Investment costs for surface drai)age systems would typically average . i500-3000/ha (less than US5 200). This is estimated to make such investment economically viable on between 5-10 million ha (Smedema, 1990). Additionally, more expensive groundwater drainage and land reclamation measures are needed in parts of northern India (para 1.36). 1.50 New Investment in Surface _Iigation. New surface irrigation projects have lower economic returns than investments to improve the producdvity of existing schemes. Typical estimates of ERRs at appraisal have been between 10% and 20%, but as discussed above, these have seldom been attained. While better project preparation and implementation can greatly improve the past low returns to new irrigation investment, all indicators point to modest economic viability. Experience suggests that moderateiy well performing, new surface irrigation projects would yield ERRs oV between 5% and 10%, while excellent performers or projects with particularly advantageous features could yield i1-15%. 1.51 Higher rates of return can generally be obtained from investments to complete the substantial backlog of ongoing irrigation projects. Acco:ding to government estimates, some 1200 major, medium and minor irrigation projects are in Lhis category, including 142 major and 216 19 See 'Uttar Pradesh Groundwater Development, Issues and Op,r,ns," India Agriculture Operations Division, World Bank, February 1991.  12 medium projects.20 Intrinsically higher viability is due to the presence of sunk costs and, hence, the higher returns of incremental investment to complete these works. However, ERRs above the opportunity cost of capital are by no means automatic. A number of past projects with poor economic returns as re-estimated at completion have been of this nature. Implementation performance remains critical, and within the substantial backlog of projects to choose from, careful selection is needed of those with the highest viability and justification. F. Needs for Future Investment Strategy Shift 1.52 Despite limited performance, expansion of irrigated agriculture has largely driven agricultural growth in India. Irrigated agriculture will remain the main source of growth for the foreseeable future, but the investment emphasis must shift from physical expansion of new capacity to productivity increases based on better performance from existing surface irrigation and selective investment in groundwater, completion of ongoing viable projects, and drainage. 1.53 An "extensive" or "protective" strategy has dominated Indian irrigation, originating from the country's experience with droughts and famines in the 19th and early 20th centuries, and more recently influenced by equity concerns and political objectives of spreading benefits as widely as possible. The goals have been to spread water over as large an area as possible and to ensure at least a basic cereal crop in the monsoon season. In practice, inequities in water distribution and the taking of water by head-end farmers has resulted in de facto intensive irrigation on sub-areas of commands, while other areas receive little or no water. Also, there is an increasingly poor fit between the water requirements (timing, reliability and quantity) of more intensive and diversified agriculture and the basic and unreliable supplies of many surface irrigation sysvems. The outcome is strained management and reduced productivity. The strategy has worked well in the northwest due to good practical application, favorable local conditions, and the additional flexibility provided by conjunctive groundwater irrigation. Elsewhere it has often encouraged use of overly optimistic design assumptions on water availability and irrigation efficiency. 1.54 The extensive strategy has also influenced the emphasis on investment for new irrigation rather than consolidation of benefits from existing investment. Since the productivity of irrigation is a more important determinant of agricultural growth than its rate of expansion (Vol U, Chapter IV), actions should concentrate on productivity of existing infrastructure. In turn, improved productivity would provide an economic justification for investment in expansion of irrigated area. Investment Choices 1.55 Emnhasis on Performance. There is substantial room for improved performance. Serious implementation issues need to be tackled, as discussed in subsequent chapters. An essential requirement is to ensure that existing irrigation remains in good repair, preserving its capability to perform. Beyond this, the most pressing need is widespread improvement in water management, complemented by improved agricultural extension on optimal water usage and agricultural practices. On most commands, the development of micionetworks is also an important call on investment, in essence, to complete utilization and magnify returns at. the local level. These thrusts can be supplemented on selected commands by higher-tech investment to further intensify production. 1.56 Selective Expansion. Expansion of irrigated area, while having smaller impact, is still .necessary to ensure agricultural growth, bui expansion needs to be refocussed on economic viability, rapid implementation, and lowest public expenditure. Expansion of private groundwater 20 Defined respectively as: major, over 10,000 ha; medium, 2(K*-10,000 ha; and minor, less than 2000 ha  13 irrigation is particularly attractive and should bc fostered through supporting public sector investments and policies. Through the medium term, expansion of surface irrigation should be almost exclusively confined to completion of viable ongoing projects. Huge scope exists and in many states completion will take over a decade. Finally, surface drainage offers opportunity for greatly improving productivity of both irrigated and rainfed agriculture in waterlogged areas. More costly drainage investments are required in some areas to preserve or rehabilitate lands affected by rising water tables and salinization. 1.57 New Surface Investment. With rare exception, there is no justification in the medium term for new surface irrigation investments. Inherent economic viability, even under good implementation, is typically modest to low. The only compelling justification for new investment is regional development concerns: where water needs are paramount and _ zher water development options are not available, such as the Narmada program for combined household, industrial, power and irrigation development in water-deficit western states, or for river basin transfers to water- short southern states where technically and economically viable. 1.58 New surface irrigation investments will become more relevant over the longer term, from early next century in some states. Expansion due to project completion will have been accomplished and prospects for groundwater development gradually exhausted (later in parts of eastern and central India) and increasingly dependent on conjunctive use, replenished by percolation from surface irrigation. Options with lower viability will then be a consideration. 1.59 Broader Economic An". A question of increasing relevance will be: What is an acceptable economic rate of return for such investments? Broader considerations than the benefits included in standard ERR analysis are probably needed. Infrastructural investments have development impact beyond first-round benefits captured by the rate-of-return calculation. They enable other higher-viability investments and multiplier effects on employment and general development, as illustrated by the impact of investment in irrigation in northwest India. In the case of well designed and implemented projects, secondary and tertiary effects can reasonably be expected and investment justified, particularly where ERRs approach the opportunity cost of capital. Clearly though, in the medium term, they have lower priority, given the more viable options for improving productivity or expanding area, with faster impact and lower public expenditure cost. Except where strong regional development considerations apply, new investments can be considered later.  14 G. Framework for Strategic Choice 1.60 Coming to grips with the inefficiencies in performance of the Indian irrigation sector necessitates answers to four straightforward questions: What is possible and how should this be planned? What are the priorities for public expenditure? How can technical implementation be improved? Who should manage and execute in order to get the best results? Answers to these deceptively simple questions will comprise the sector's strategic framework for the 1990s. In concert, the answers will guide policymakers and imtrlemeritors on actions to improve productivity, ensure sustainability, refocus investment, and strengtheti sectoral management. Four major actions are required: 1.61 Forge a Coherent WYater Resources Policy. Complete data on the availability and distribution of water are needed as a first-step in planning investment for effective usage of a limited resource. The natural hydrological unit, the river basin, is the essential planning unit, yet its boundaries rarely coincide with political bouindaries. Ground rules and mechanisms for sharing water and coordinating usage across boundaries and among competing users -- agriculture, industry, municipalities, etc. -- need to be hammered out. Holistic planning and management are needed as demands for water-sharing increase. With this approach, firmer grounds are established for investment choice that is responsive to users, sectors, regions and technology. Environmental and resettlement policies also need improvement. 1.62 F_dri_tize Expenditure and Tighten Financial Management. Public funds for investment in irrigation are limited by budgetary constraints and further are being squeezed by recurrent expenditure. More transparent budgeting and accounting procedures would point to shrinkage in investmnent per se. The widening discrepancy between expenditure and revenue is rooted in extremes on both sides of the ledger -- rapidly rising staff costs and declining revenue from user charges. Both must be brought under control. Investment must be prioritized to achieve maximum impact. This context envelopes decisions on whether to upgrade existing infrastrucrure or build new capacity and on what balance to strike among groundwater, surface irrigation, and drainage investment, to cite a few. 1.63 Improve Technical Performance. Improving sector perfornance hinges on more than narrow technical choices. Broadly, it encompasses strategy, agroclimate, project design, construction, operation, maintenance, management, and responsiveness to community needs. The overall scope for improvement on all these fronts and in most states is considerable. In particular, technical implementation requires improvement in water management, maintenance, design and construction. 1.64 Reorder overnment/Private Roles and Build Capacity. Investment in irrigation does not necessitate exclusive public management and investment. Government cannot afford to undertake and manage all irrigation investment. Nor should it. Critical analysis is needed on what government should be doing and where privatization would be beneficial. Autonomous commands, farmer operated and maintained micronetworks, and consultant, NGO and university involvement in data collection, analysis, project planning and design are only a few of the possibilities. Where govemment does plan and execute, its capacity to do so must be enhanced and its units restructured to accommodate more specialist expertise across more disciplines. The construction engineering approach should give way to a broader-spectrum management culture. 1.65 The chapters that follow will address these four strategic actions in turn. (See Annex I for matrix table of highlights of recommended action.)  15 CHAPTER II. WATER RESOURCE POLICY AND PLANNING 2.1 Effective planning and coordination of water resource development are essential if India's scarce water resources are to be efficiently developed. GOI's National Water Policy (1987) provides a framework for coordinated water development across states and alternative uses of water, and it emphasizes the need for river basin planning. However, the National Water Policy has had little operational impact due to lack of institutional mechanisms to plan, coordinate and implement water development across state boundaries and among users (municipal and industrial, irrigation, hydroelectric, etc.). A complication in this regard relates to the federal government structure in India and the status of water in the constitution as a "state subject". Yet the water resource is defined by the natural hydrological unit, the river basin, and only one of India's 18 major river basins is contained in a single state. Another problem is limited data and mechanisms for sharing of data between states and between users within states. Water pricing and rationing can also be used more effectively for efficient allocation and use of water. As water scarcity increases and multiple demands on usage grow, conflicts among users will increase. Already, such conflicts are common. More important, uncoordinated past and present decisions on water use are diminishing possibilities for optimal water use and will limit future development prospects.21 A. Situation and Trends 2.2 Development Strategy. Maximizing returns to water implies widespread infrastructure to spread water. India's historic irrigation development strategy of spreading water as widely as possible ("protective" or "eytensive" irrigation) has resulted, in practice, in substantial shortfalls, mainli due to design and implementation problems, except in the northwest. Stretching water suppl.v s over larger areas to address protective or equity concerns cannot be applied indefinitely withoiw. eventual conflict with efficiency and production objectives. Beyond a certain point, econcmnih viability diminishes as net incremental benefits fail to cover the additional investment costs in distribution networks. Eventually, gross output will also fall as water quantities become so overf ietched that productive impact is impaired or farmers, frustrated by limited water, take more th< . their share or break the system. Overly optimistic design assumptions concerning availabiibtk of water and feasible irrigation efficiency have also meant that many systems do not have enoul-, water to operate as assumed at the planning stage. 2.3 Li.iiitcd Availability. India's water resources are limited in aggregate and distributed unequally across regions. Both groundwater and surface irrigation development are constrained by such physical limits and in some regions these limits have already been reached. In the northwest, water development has nearly reached its physical limit. In Tamil Nadu, state water resources have been fully exploited and further development will depend on greater access to water from the Cauvery river fiowing through upstream Karnataka, or to any eventual prospect for transferring water across the western Ghats from Kerala. In western India further development depends on major storages-based schemes (e.g., Narmada). Development is also constrained by riparian rights of other countries. The successful Indus Basin Treaty between India and Pakistan has clarified the respective usage rights of the two countries and permitted non-conflicting investment, but its adherence is part of the limit on northwest water development. For the Ganges basin, dry season flow requirements to Farakka ba. age serving downstream Bangladesh will need careful planning of future water development on India's part. The interlinked nature of groundwater and surface water must also be recognized. Exploitation of one affects the potential development of the other, yet each is measured independently and water development for these resources has seldom been planned conjunctively. 21 Background on tth, status of existing knowledge is contained in "Assessment of Indian Water Resources and Irigation Development Poiential" (W. Barber, C.C. Patel et al, Irrigation Sector Review Background Paper, 1990).  16 2.4 Further, as a scarce resource relative to land, efficiency in water use must be encouraged, thi ugh rationing or pricing. Apant fromil thte warabandi system (rationing) and use of diesel fuel for pumps (with fuel costs incurTed in piopo)lion to the quantity of water used), measures to encourage economy of water use are seldoom applied. Crop-based water charges on surface schemes (high water charges for more wate intensive crops), to some extent capttre differences in use by crop, but not the actual water use. Further, they are set too low to have any impact on cropping decisions. This lack of effective rationing or water pricing has its consequences on water- use efficiency. Sugar growers in Maharashtra occupy 10% of cropped area and use 50% of irrigation water (Rath and Mitra, 1987). Abundant water in some commands of southern India, in designated "wet" areas of localization schemes, and in head-reaches of most commands throughout India implies lack of water for other farmers. 2.5 _ppetrin~ gsf Wa te. Community water for urban and rural populations, livestock, and water for industry, thermal power, hydrop(ower, and river flushing and navigation compete with irrigation. Irrigation is the largest user of water in India, accounting for 93% of present gross water use. However, community use is expected to double and industrial and power use increase seven-fold by 2025. While this would not limit the availability of water for irrigation overall (irrigation would still use an estimated 90% of overall net demand), the regional distribution of water demand relative to supply is a problem. High community and industrial demand already present water supply problems around urban concentrations in Delhi, Madras, Bombay and other cities, particularly in peninsular India. The concentration of heavy industry and thermal power generation around Jamshedpur in Bihar will take the bulk of Bihar's share of future water development under the Subernarekha development program. Areas in some southern states (Tamil Nadu) and western states (Gujarat, Rajasthan) are experiencing shortages in community and industrial supplies. Water supplies for drinking and community use must be secure and take evident precedence over irrigation. Irrigation development will thus have to be planned and implemented in the context of multi-purpose usage and tne usually higher claim on water for municipal, industrial and some ither uses. Planning is woefully inadequate and under- emphasized. Initiatives to improve matters have been beset by the political difficulties of reaching agreement between states. 2.6 Data Collection and Sharng. For most river basins, data collection facilities provide an incomplete basis for measuring water resources, monitoring water usage and planning development. Further, data is collected disparately by different agencies. As important, such data is not always published or readily accessible; both between states and even between departments and agencies within states. These problems have been emphasized by MOWR and state governments, but improvement actions are lagging. Proposed measures include: further development of data collection facilities covering surface water, groundwater and water quality; creation of data banks; improving interlinkages for data collection and sharing between agencies and states; and improved water resources analysis on a river basin basis, including combined analysis of surface and groundwater resources and usage.22 2.7 Institutional Constraints andMechanisrms. As situations of water scarcity and competing needs mount, inadequacies in coordination of water planning and usage have become increasingly apparent. The constitutional nature of water as a "state subject" means that states have full authority over water use and legislation within their borders, encouraging disparate decisions without reference to river basin usage as a whole. Two acts have provided some mechanisms to tackle inter-state water issues. The River Boards Act of 1956 gives GOI the authority to set up a river board for any inter-state river, but the boards' functions are solely advisory. More significant powers are available under the Inter-State Water Disputes Act of 1956, which authorizes GOI to refer any water dispute to a legal tribunal for adjudication, with findings of the tribunal being final. 22 As recommended in various govemment reports; foi instance, preparation reports for a National Hydrology Project prepared by Rourkee Institute, CWC and CGWB (1987-89) under guidance of the High Level Govemment Committee appointed for this purpose.  1'/ For groun3dwater development, state g(o)\x'i mlJ i-iis hlS ., *:t\ lii('l.i'i o( -nstitutional authority. While some success in hartnonizib, intcr-siditc il)lial ('vJioplci)ait lias been achlieved through cooperation anid specific tribunials ( for instaince, in tlih shmarinf of Namrmada, Krishna and Subernareklha waters according to Lilibiual awards), thl miore comrmon situation is lack of agreement. The findings of the Ravi Beas tribunal in the Indus basin have yet to be finalized, and a highly contentious inter-state dispute aniniag Tainil Nadu, Kaxnataka and Kerala regarding the waters of the Cauvery lhas remained unrcsolved siince 1974. Shaiinig of Yamuna waters between Haryana and lJttar lradesh is still a matter of di ,pute, despito an inter state agreement. 2.8 a Ri Basi P-lajinigg. In indusuial countries where water-sharing issues have persisted for some time, river basini planning has usually become standard practice with holistic consideration of its mnultiple uses.23 Progress towards holistic water planning and implementation is iiore variable in Asia. In Cthinla, liver basin planning is fairly well established. In the Philippines, the National Water Resources Board has the mandate to conduct river basin surveys and coordinate water allocatio<n wnong useis. 'the couinthy is divided iinto water resource regionis partly coricespoidinig to i iver Iasilns bu1t also infilulnce-d by administrative boundaries. In Indonesia, where water is constitutionally iecognized as a centr-al government as well as local responsibility, iveir basin plannting has begun, ihouigh it is still rudimentary. Three river basin implementation authorities have been established, though their roles are not yet clearly defined. In India, the Darnodar Valley representedl a seCotis attempt at plamniiig and implementing whole basin development. Elsewhere, ad hoc decisions have generally been necessary, usually project by project, with high opportunity costs in term-ls o' conflicting developrnent or lost opportunities for better resource-sharing.24 2.9 Watershed DeveloDmenit. Aniother iieed commented on by the Ministries of Water Resources, Agriculture and Environmryent and Forests is for liver basin planning to encompass watershed development as a whole. This includes rainifed agliculture, forests and pasture, and techniques for water and soil conseivationi; for instance, on-farm vetiver hedges, contour ploughing, forestation programs and groulndwater recharge. (para 4.17) by surface irrigation networks.25 A current government conceirn is to create awareiiess and institutional capabilities to plan and implerment land and water developmnent in this much broader sense. This would include consideration of the whole water recharge and run-off process for better management of groundwater levels and river flows, and integrated planning for all uses of water (non-agricultural and agricultural). 23 In England and Wales, river basin water resource authorities are separate from local administrations and have overriding powers as concerns water administration following a principle established in the 16th century. Similar procedures were initiated in France in the 19th century, and completed in the 1960s through the creation of river basin agencies for its main rivers. Canada and Germany also follow a 6iver basin approach. In the United States, federal authority over water is restricted to a consultative process including examination of environmental issues. But major interstate rivers have come under river basin planning commissions, and agreements between states are bound by inter-state compacts recognized under law. The legislative process and commercial water transactions are principal means for resolving water disputes, a procedure less suited for countries where the legislative apparatus and water pricing are less established. Unresolved disputes are common in the United States, suggesting that greater central capability to settle disputes (as in the United Kingdom and France) may be more appropriate, especially in developing countries. 24 For fiurher discussion of water planning and resource issues, see Asia Water Resources Study (Asia Region, Agricultural Technical Division, Wcrld Bank, draft, 1991). 25 Such concerns are also discussed in a niumnber of World Baia. repcnts on India, for instance: the reviews of "Rainfed AgriculUre and Watershed Development", and "Wasteland Oevelopment" (both India Department, 1988); and "Vetiver Grass, A Method of Vegetative Soil and Moisture Conservation" (Grimshaw and Greenfield, 1987). Four watershed developmcnt projects -- Pilot Rainfed (Cr. 1424, FY84), Himalayan Watershed (Ln.2295, FY83), National Watershed Development (Hills) (Ln.3175, FY90) and (Plains) (l.n.3197, FY90) -- encompass this approach. Refer also to discussion of environmental issues in Chapter IV.  18 2.10 Coordination Needs. In India, neither state nor central govemment admiinistrations has been able to effectively deal with the whole river basin as the planning unit. The states have specific interests which are generally narrower than those required for whole basin planning. The center has limited powers and would in any case be too far removed for the detailed analysis and decision-making required. Institutional arrangements and lack of coordinating mechanisms complicate policy and implementation at both levels. Separate GOI agencies exist for surface water (the Central Water Commission, CWC) and for groundwater (the Central Groundwater Board, CGWB). The results are separaite accounting of each and over-estimation of combined water resources.26 Efforts to coordinate activities by the Ministry of Water Resources (MOWR) and other central ministries involved with water (mainly, Energy for hydro and thermal power, Environment and Forests for catchment conservation and pollution, Urban Development for urban water supply and sanitation, and the Planning Commission) have suffered from limited powers at the central level and the difficulties inherent in cooperating across sectors. Cooperation among departments responsible for different aspects of water use at the state level has also proven difficult. B. Tackling Resource Planning and Policy Issues 2.11 River Basin Planning. Establishment of procedures and institutional capabilities to undertake holistic planning and coordination of water development by river basin and including all uses of water is the most essential need. This requires improved data collection and analysis, and the creation of coordinating institutions at river basin and sub-basin levels. MOWR is coordinating state and central government discussion towards this objective through its National Water Board (NWB) established in September, 1990. Implementation of river basin planning is currently seen by the NWB as an evolutionary process involving the following steps: (i) improved collection and sharing of river basin data; (ii) establishment of river basin planning boards; (iii) joint ventures for inter-state development projects; and (iv) possible whole-basin management over time. 2.12 Inter-State Cooperation. Rapid progress towards river basin planning in China, Indonesia and the Philippines has been enabled by the power of central government to lead the decision- making process, which at least one regional or sectoral entity would otherwise obstruct (the one which might stand to lose resource usage). This is less easy in the Indian federal structure where state rights are greater. An option would be to make water resource planning a "concurrent" subject (joint responsibility of both central and state governments), a situation found for other sectors (e.g, power, national highways), which arguably require no more inter-state and state- central cooperation than water. While implementation at river basin, sub-basin and project levels would remain a local responsibility, greater prospects would be present to effect change and resolve conflicts. Such a constitutional amendment would, however, require two-thirds majority agreement by states and appears unlikely. Instead, MOWR and state governments are in favor of progressively implementing institutional interlinkages to improve inter-state cooperation. In this process, it is foreseen that MOWR can play an important facilitating role. The tribunal awards system can also be used more systematically. 2.13 Central Coordination. The present separation of surface and groundwater responsibilities between CWC and CGWB encourages disparate consideration of surface and groundwater resources. The two agencies should be merged or better mechanisms created through MOWR for a joint approach. Also, there is a need for GOI to provide greater cross-sectoral and inter-state coordination for water planning, a role that could be performed by the recently established National Water Board (para 2.11). 26 As commented upon by CWC (1988), but not yet adjusted for in official estimates of water availability (parm 1.29).  19 2.14 Water Pricing. As an allocative mechanism, water pricing should increasingly be used, including in agreements on water-sharing. Agreement between states and different users in India would be facilitated if water was priced. It is employed, inadequately, by states for distribution to nonagricultural users, but not between states. Yet doing so is a likely precondition for successful future river basin transfer schemes. For instance, any future plan for transfer of water from the westem ghats in Kerala to water-short Tamil Nadu could involve fees to Kerala for the opportunity coDst (e.g., foregone future hydroelectric power) and investments involved. Similarly, agreement betveen India and Nepal for investment in storage in Nepal's part of the Gangetic basin would revjuire clear agreement on water-sharing, investment funding and fees charged for water and hydroelectric power. C. Actions to Improve Water Policy and Planning 2.15 The following actions are recommended: (a) Reaffrmn and apply the National Water Policy. Each state should reaffirm the National Water Policy and prepare a "State Water Policy" fitting within the framework of that policy and river basin plans.27 The National Water Policy paper may require some updating as part of this process. Central Plan funding for state irrigation investments should be made conditional on the existence of a state water policy and its fit within the river basin plans; (b) Establish river basin planning commissions for major river basins. These should be charged with preparing river basin plans for multiple use of water and for coordinating and authorizing speaific developments; (c) Establish autonomous implementation authqrities for specific projects r iring coordinated water use between states or users. These autonomous agencies should operate within the general guidance of river basin commissions to develop and manage water28; (d) Improve data collection, resource analysis and data sharing. Data collection, monitoring and analysis by central institutions and state governments need to be enhanced for better planning (including combined assessment of surface and groundwater :resources and monitoring of environmental impacts). This needs to include establishment of better mechanisms to pool data and make data generally available (from different departments within states and for data sharing between states and between states and the center); 27 While the existing z o,n is that probably no state has a state water policy in the full detail recommended here, a number of states have made initiatives that could be capitalized on in development of a full State Water Policy. Amongst such initiatives, for instance, is a draft state water policy prepared by Madhya Pradesh, the ongoing work of the Second Irrigation Commission in Bihar and the wock of the Standing Irrigation and Water Resource Commission in Tamil Nadu. 28 The Bhakra-Beas Management Board is a useful example in this direction. It coordinates decisions among the implementing states (Punjab, Haryana and Rajasthan), and has been largely successful in this function. At least similar institutions are needed for many other major schemes as they frequently involve significant multiple use or more than one state. For instance, in the Narnada program, the present semi-autonomous institution (the Nannada Control Authority) has proven useful for inter-state coordination of planning but will need an eventual expanded role and grar authority over scheme operation and water use. Likewise, Subernarekha waters, shared among three states and with large municipal and industrial use, will require a unified implementation agency for coordinated scheme managemenL In cases where the scheme embraces the entire basin, the river basin commission can itself take on the implementation role.  20 (e) Use pricing to allocate water. To generate funds and facilitate inter-basin transfers, guide municipal and industrial use, reduce overexploitation of groundwater and electricity and, as practical, guide surface irrigation use; (f) Impr centr coordinatmn. Improve inter-state and state-central coordination of water planning through the National Water Board (para 2.11); and combine, or at least better coordinate, the roles of CWC and CGWB; and (g) 3ffecdyy use rIdbuna! Awards. Greater recourse should be made to Legal Tribunals, as allowed for under the Inter-State Water Disputes Act of 1956 (para 2.7), to adjudicate current or future water-sharing disputes. Central Plan funding for a water development project should be conditional on satisfactory resolution of water-sharing issues.  21 CHAER m. EXPENDlTURE AND FINANCIAL MANAGEMENT A. Investment Focus Situation and Trends 3.1 Investment in irrigation will not by itself ensure good agricultural growth. To be effective, it must be focussed on the right priorities and implemented through the right project activity. In India, scarce government resources have not been getting to the most urgent and highest-return priorities in the irrigation sector, and inadequate use is being made of private capital for irrigation development. Substantial wastage is occurring. Public investment is also under pressure as a result of declining revenues from water charges. Radical improvement in the investment focus and financial management of irrigation is required. The only tenable choice for government is to engage in a bold and rigorous effort to prioritize investment and tighten management, while at the same time, addressing key performance problems. Expenditure Priorities 3.2 Government should focus expenditures on investment that will enable irrigation to make a maximum and sustainable impact on agricultural growth. Past expenditures have, in many ways, been the opposite of this. Public funding has emphasized new, long-gestation surface irrigation, and has resulted in a proliferation of incomplete schemes. More viable options, including encouragement of private investment (private rather than public tubewells, farmer rather than government maintenance and operation of micronetworks), have been neglected. Obvious and overriding priorities are improved water management and maintenance. Expansion of irrigated area will also be necessary to improve agricultural growth. Investments with high economic viability, productive impact in as short a time as possible, and relatively low public cost per hectare are key. Additionally, project preparation and selection must be based on broader data collection and rigorous analysis, and a pipeline of quality projects developed to ensure selectivity and more timely and better implementation. The following should be financed: (a) Maintenance. Maintenance cannot be neglected and must receive first call on govemment expenditures. The past investment in over 16 million ha of surface irrigation is increasingly falling into disrepair due to inadequate maintenance in nearly all states. Preserving the existing investment is essential or over a century of effort will return to rainfed conditions. (b) Productivity Improvements in Existing Schemes. The critical need is to improve deficient water management. Achieving impact as quickly as possible necessitates a low-cost approach, focussed on improved seasonal planning and operation of surface schemes, complemented by selective investment on key control structures and command communications. This can be supplemented by a more limited program to modernize selected commands, backed by technology research and initial piloting. In parallel, extension for inigated agriculture also needs improvement. (c) Supp?ort private groundwater development This has been the largest contributor to irrigation expauision and improvement over the past two decades. Scope is particularly large for further investment in eastern India. Government should enhance its support for private investment through: credit and (where justifiable) investment subsidies for poor farmers; expansion of rural electrification but concomitant removal of electricity subsidies and introduction of meter-based pricing; recharge programs for areas with falling water tables; research and promotion of energy-efficient pumps; and monitoring groundwater levels and any necessary regulatory functions.  22 (d) Completion of viable. ongoing surface irrigation projects. Surface irrigation investment should be restricted to completing the substantial backlog of incomplete projects. The government estimates that incomplete projects represent 12.5 million ha of potential and completion costs of Rs 240 billion (US$13 billion). Although estimates of the backlog, the area involved, and the costs of completion vary, the important need is to select amongst the projects that are incomplete and prioritize expenditure to those completions with highest viability and prospects for rapid implementation. Such completions would involve lower incremental costs per hectare, higher viability, and shorter implementation periods than new projects.29 In some states completion of the backlog would take at least a decade. No new investments should be undertaken in a state until viable ongoing projects are completed. (e) Supprt Farmer Maintenance and Development of Micronetworks. Investment in water courses, field channels and field drains is required on most commands, even those considered "complete" in terms of main infrastructu .. While this need is well recognised, there has been an increasing tendency for the public sector to construct and maintain such networks. This is unaffordable and leads to perpetual reliance on the state. A drive should be implemented in each state to hand over responsibility for maintenance of micronetworks to farmers and progressively increase their role in construction. Improved water management and maintenance to the government outlet needs to be accompanied by an upgraded technical assistance and extension service to help farmers operate and maintain micronetworks and establish water user associations (para 5.46), plus provision of credit and perhaps matching investment grants for micronetwork upgrading, rehabilitation or construction. The progression towards farmer responsibility in rnicronetwork construction or improvement may take longer to achieve as more experience with this approach is required at both ID and farmer levels. As a minimum, each state should introduce active participation by farmers in micronetwork design, some labor contribution, and (even if quite nominal) a financial contribution. This would help create a sense of farmer ownership and responsibility for subsequent maintenance. (f) Drainage. flood control and land rehabilitation. Investment in surface drainage, previously neglected, is now needed in parts of many commands, including more expensive groundwater drainage in some areas affected by rising water tables and salinization. The potential aggregate area where surface drainage would be economically viable is some 5-10 million ha. In many areas drainage investment would have greater productive impact, and at much lower cost, than investment in surface imigation. (g) Storage and river basin transfers. The above items take precedence over new irrigation schemes. However, over the longer term, investment in water storage will increasingly be required to capture seasonally concentrated rainfall and to respond to crop requirements outside the monsoon season. In some water-short areas, even now, selective new investment in storage-based irrigation, usually associated with other uses of water, is justified (e.g., the Narmada development program for western India). Also, where technically and economically viable, inter-basin water transfers serving water-short southeastern India may need to be considered. 29 To sharpen the list of candidate projects subject to technical and economic viabiity assessment for completion of works, G01 has suggested that initial selection be confined to projects where at least 25% of intended investment has already occuwred.  23 (h) Rehabilitation. Regrettably, reinvestment in infrastructure that has deteriorated due to poor construction and maintenance is necessary. In a number of states, this need will demand a major p&tion of public expenditures on irrigation and will reduce funds for expansion for many years to come. 3.3 A consensus has been established with GOI and at the level of most states on the priorities discussed above. However, the challenge will be to implement this expenditure focus and, at state levels, not to undermine expenditure priorities as a result of political pressures for investment in new areas, or diversion of available funds on unnecessary staffing and financial wastage. For instance, the Seventh Plan stressed project completions and, with lesser emphasis, maintenance; yet impact, especially regarding maintenance which has declined (Chapter IV), has been disappointing. To this end, GOI intends to require detailed development and expenditure plans and reviews by each state responsive to their priorities. This is beginning in the preparation exercise for the Eighth Five-Year Plan (1992-1997). Additionally, stricter annual monitoring of state expenditues is required by MOWR, the Planning Commission, and the Finance Commission (para 3.21). G01 has also recommended consideration of use of Indian academic centers to do part of the analysis for state Plan priorities and annual monitoring, as a means of enlisting the benefits of independent external advice. Individual state expenditure programs will need tailoring to the specific needs of the state concerned.30 However, most of these priorities -- especially maintenance and productivity improvements, but also micronetwork development and viable short- term project completions -- have universal priority for government expenditure. Prject Preparation and Selection 3.4 Pressures to invest in irrigation have led government to sacrifice the quality of technical preparation, financing investments without full project preparation based on adequate hydrological, technical, agronomic and socioeconomic data, and without a detailed and costed component-by- component technical design. Serious deficiencies in economic evaluation and technical preparation are the cause of many of India's problems with irrigation implementation. Once priorities for investment are in place, choosing the right project is paramount. Project identification, preparalion and appraisal procedures, as well as the criteria and processes used in project selection, need overhauling. Unlike industrial and other infrastructural sectors in India, where calculation of an economic rate of return (ERR) has become standard procedure, irrigation investments are evaluated on the basis of undiscounted benefit/cost ratios. The cut-off benefit/cost ratio for irrigation is 1.':1, with a ratio greater than 1:1 accepted in drought-prone areas. Prices used are financial and sensitivity analysis is rarely performed. These ratios are calculated at the end of project preparation, rather than as a decision-making aid during project identification and when evaluating design alternatives. A common practice is to ensure project acceptance by inflating benefits and underestimating costs. Despite a lengthy multi-stage clearance process, involving numerous government departments at state and central levels, scrutiny of key technical assumptions is often weak. Economic and financial evaluations are also weak, due to the critical absence in state irrigation departments and in the CWC of staff qualified in these disciplines. Finally, a serious consequence of these deficiencies is a lack of well prepared projects ready for appraisal. A pipeline 30 Punjab, for instance, with irrigation virtually fully developed and usually in good operating order, needs continued '.nphasis on assuring adequate maintenance funds, some improvements to existing iigation structures, and marginal extensions where feasible, plus a large investment program in drainage to combat rising groundwater in some areas. St&aes such as Madhya Pradesh, have large scope over several decades for completing those incomplete projects that would be viable, but as first caus on annual expenditure need to assure good maintenance and improved operadons of existing schemes, including rehabilitation of deteriorated schemes. In much of eastern India, overwhelming priorities are to address weak perfomuance and deterioration of exisdng irrigation - adequate maintenance and improved operations, substanti rehabilitation and micronetwork development - diminishing revenues for expansion other than thiough private groundwater development. Individual states may have special additional needs; for instance, the sodic lands problem in parts of Uttar Pradesh, and tackling salinization and sea erosion in Kerala (refer Volume II, Chapter H and Annex H for additional state-by-state discussion).  24 of irrigation projects needs to be developed so that selection of investment will be based on quality. Then, there is need for an objective selection process, with prominence placed on economic viability. Aions tQ Inxroe Project Inyvestmnent 3.5 The following actions are recommended: (a) Create a pipline -of hort- and meium-term irrigation p osals, focussing on priority options identified above. (b) Useore rigorous guidelines for technical preparation of irrigation and drainage investments; (c) Revise the methodology for economic evaluation of irrigation projects based on standard discounted cost/benefit criteria (ERR or discounted net benefit' cost ratio). A committee formned by the Planning Commission is currently reviewing economic evaluation procedures. A number of changes, including use of discounting, are likely to be introduced in the committee's final report; (d) Upgrade the proiect preparation and appraisal canability in each state irrigation department and at the level of CWC, involving more thorough technical analysis, funding of physical and socioeconomic data collection, staff training and use of consultants; (e) Establish or. where existing. upgrade project monitoring and evaluation units in state inigation departments and CWC; (f) Contract part of monitoring and evaluation work and impact evaluation to consultants and universities; and (g) Engage local and selected national universities in is d m and evaluation studies, to increase the public accountability of decision-making and performance. B. Financial Management 3.6 Until the early 1950s, irrigation made a net positive contribution to government finances. Revenues from water charges exceeded government expenditures for operations and maintenance plus imputed interest on investment. Now the sector is a financial burden. By 1967/68, capital costs were no longer beirng recovered and a net annual loss of Rs 580 million was incurred (on revenues from water charges less recurrent expenditures). Deterioration accelerated in the 1980s. By 1988/89, current expenditures on operations and maintenance on major and medium irrigation projects exceeded revenues from water charges by Rs 23.5 billion.31 Rural electricity subsidies, primarily for pumping water from tubewells, accounted for another Rs 14.6 billion (1986/87, current prices). Subsidies to irrigation grew by 10% per annum in the 1980s and rural electricity subsidies grew by 15% per annum. Irrigation and electricity subsidies now represent the bulk of rumral subsidies in India32; in 1986/87, irrigation was 37% and electricity 27% of total rural subsidies. 31 Further, surface irrigation subsidies have been expressed here without inclusion of capital cost recovery. 32 Source: "Agriculture: Challenges and Opportunities" (India 1991 Country Economic Memorandum, Agriculture Operations Division, India Department, World Bank).  25 3.7 In the 1980s, while current expenditurcs increased in real terms by 280%, revenues from water charges increased by only 29%, despite expansion of irrigated area (Table 3.1). Current expenditures for major and medium irrigation projects are now 14 times revenues, representing 6% of total state current expenditures, while water charges are 0.4% of total state revenues. At the same time, total expenditures on irrigation have shifted from capital investment (irrigation expansion) to recurrent costs. In 1980/81, 78% of expenditures went to investmnents; by 1988/89, the share of investments had fallen to 55%.33 Costs per hectare irrigated for new irrigation have also mounted. Table 3 .1: Revenues and Expenditures On Major & Medium Irrigation Projects (Rs Billion in Constant 1980/81 Prices) 1980/81 1288/89 Current Revenues (water charges) 0.9 1.1 (+29%) at Current Expenditures (non-Plan) 3.9 15.0 (+280%) a/ Capital Expenditures (Plan disbursements) 14.0 18.3 (+31%) A/ Current Revenues as % of Current Expenditures 22% 7.5% Capital (Plan) outlays as % of Total Expenditures (Plan and non-Plan) 78% 55% M/ Percent increase of 1988/89 over 1980/81. Source: Adapted from Reserve Barnk of India Bulletins. 3.8 Rising Current Exnditures. The steep rise in current expenditures has been primarily due to growth in irrigation department wage bills, mostly due to burgeoning staff numbers. Standard state data do not systemically disaggregate non-Plan expenditures into component parts (staff numbers, wage increases, works, etc.), but data from individual studies in selected states and commands may illustrate the situation applicable in varying degree in nearly all states. In some states and commands analyzed, current expenditures in the early 1980s more than doubled in four years, almost entirely due to increasing staff costs.34 3.9 Inadequate Funding of Maintenance. Little productive impact has been achieved from the escalation of current expenditures. Instead of improving operations and maintenance, urgently needed in most states, staff numbers and wage bills have absorbed most incremental expenditure. Even in Punjab, where maintenance has been comparatively good, an average of 85% of the non- 33 Financing in India is subdivided between "Plan" funding and "non-Plan" funding. Plan funds are for investment purposes (new investments, completion of ongoing investments, rehabilitation, etc.). while non-Plan funds go to current expenditures such as, in irrigation, operations and maintenance including staff salaries. Plan funds allocated in the states have two components; funds allocated to the states by central govermment including foreign aid contributions, and funds originating from state exchequer. The center also makes general financial allocations to the states which supplements state non-Plan expenditures. Except where otherwise indicated, current expenditures in irrigation refer to nen-Plan funds, and investment or capital expenditues refer to Plan funds from both central and state resources. Irrigation sector revenues refer to direct revenues from irrigation in the forn of water charges paid by farmers (where reference is made to groundwater irrigation, electricity tariffs are also rdevant). 34 See analysis in Vol. II, Chapter V. An extreme situation is found in Bihar where establishment (personnel) costs increased by 148% in only four years (1980-81 to 1984-85) while expenditures on works relaued o maintenance declined slightly in the same period. The growth of establishment costs in Bihar was influenced bodh by a doubling of the costs of the revenues establishment (personnel involved with collection of water charges) and by a 2.6 fold increese in the costs for operation and maintenance staff.  26 Plan irrigation budget between 1986 and 1989 was spent on salaries and wages, and in the same period identifiable expenditures on maintenance decreased by 40%. 3.10 Increasing Investment Costs. Comparing the Fifth (1974-79) and Sixth (1980-85) Plans, investment costs per hectare increased by 57% in real terms.35 Based on documented project experience, this trend probably continued over the past five years. Part of the increase reflects the higher cost of progressively more difficult project cptions over timc, but this hardly explains the increase in so short a time. Further examination is required. Likely causes are laxity in contracting and supervision of construction, and inadequate systems for monitoring, cost analysis and control of expenditures. 3.11 Declining Revenues from Irrigation . Water charges are the rniain direct source of revenues from the irrigation sector. For surface schemes, these are levied annually on a per hectare-crop basis, with rates depending on crop and season. Rates are generally higher for more water- intensive crops and in seasons when irrigation water is more scarce. For electric tubewells, an annual fixed charge per unit of horsepower is applied. Water charges vary considerably across states, as do collection rates.36 The slow growth of revenues from water charges reflects their infrequent revision, resulting in erosion of real values over time. Water charges have not been changed since the mid-1980s in most states and not since the mid-1970s in Punjab, Haryana, West Bengal and Tamil Nadu. Compounding the problem, collection rates have declined in a number of states, especially in the last several years. Actions to buprove Financial Management 3.12 The origins of financial problems in the irrigation sector are weak financial management and monitoring at the state level. Political and rent-seeking pressures perpetuate the situation. In most states, no meaningful financial analysis, planning and monitoring are undertaken. Irrigation departments seldom contain financial analysts and there is little cost-accounting to probe what is happening to expenditures.37 Management information systems, where present, do not usually generate useful and timely data for decision-making. Availability of data on staff levels and expenditures, unit costs and cost-recovery over time is scanty, and usually out-dated. Verifying expenditures is also difficult because desk audits do not check financial statements against field actioi.s and materials used, a difficult .sk with dispersed field operations. 3.13 Substantial improvement in sectoral efficiency can be made if these issues are tackled. Conversely, without action, the sector's drain on public resources may soon reach unsustainable proportions, a situation nearly at hand in some states. 3.14 To be effective, strong political support is necessary from state leadership in providing back-up to state irrigation and finance departments, and in creating a supportive environment for change. Recommended actions at the state level are: 35 Average for 14 states based on New Concept Consultancy Services (1989) and Ravishankar (1990). 36 The collection rate is the actual money coliected expressed as a percentage of the money due based on existing water charge levels and irrigated area. 37 Some useful initiatves are now being taken. In Orissa, the Irrigation Department has obtained agreement from the state budgeting authorities to have a budgetary sub-heading for 'staff' (previously grouped with "works") in the O&M budget. This will limit line managers (e.g., executive engineers) to keeping stafing within the budget for staff and protect expenditures on materials.  27 (a) Implement an interim action program. A quick diagnostic analysis of irrigation expenditures by component over time should be undertaken by each state. This should include ccmparison of achievements based on physical indicators against expenditures. The review should focus on evolution of staff numbers and the wage bill; costs per unit of infrastructure constructed or under operation; and maintenance needs compared with actual expenditure on maintenance. Even with the limited data available, key problems and needs in any state are readily identifiable. Based on this diagnosis, an interim action program should be drawn up and implemented to tackle the most critical deficiencies: halting growth in staff numbers, providing adequate funding of maintenance, and controlling construction costs. Morm demiled and fundamental follow- up can occur over dtme. (b) Initiate cost-accounting and develop financial management skills. Staff capabilities, mechanisms for collection and analysis of data, and management reporting procedures need to be established for each command and at ID headquarters. This would provide the management information system required for taking actions to improve the efficiency of expenditures by operation (e.g., staff versus works expenditures, breakdown of materials used and costs, comparative costing on a per-hectare or kilometer-of-canal basis). Consultant assistance will usually be necessary to help establish the MIS. (c) Reduce staff numbers. A state plan to reduce irrigation sector staff numbers should be drawn up, implemented, and annually monitored (See Chapter V for alternatives to expanded governrment personnel in irrigation). A number of states have recently made a decision to halt recruitment of additional ID staff. This needs further application through systematic use of natural attridon as a result of retirements without hiring replacement staff. Together with other attrition (e.g., in-career resignations), annual ID staff reductions of 3-5% should be attainable.38 This needs to be monitored through the ID's MIS system and backed up in the state's budgetary process through provision of separate line items in the recurrent budget for staff costs and costs of works and materials. (d) Plan and fund maintenance fully. Detailed state irrigation maintenance plans and budgets following a standard format should be prepared annually. The maintenance budget should be based on the actual needs detailed on a command-by-command basis. Maintenance expenditures corresponding to the budget should receive top funding priority rather than the present last priority.39 Maintenance should be closely monitored, both in terms of financial allocations and field actions.40 38 More comprehensive staff-reduction action may be necessary in certain cases (for instance, where water charges collection, manag- ment of micronetworks and deep tubewell operation is handed over to local bodies; e.g., Panchayats, Water User Associations. Modalities, including appropriate separation benefits, would need to be carefully and sensitively worked out for such cases. 39 As commented on by MOWR, contingencies must also be allowed for in maintenance funding; for instance, higher than expected budgetary needs in any given year to repair flood damage, perhaps requiring a special augmentaion to the maintenance budget. 40 Cost nonrs per hectare have been prepared in the past by CWC to guide funding decisions for maintenance. Actual allocations by state govemments towards maintenance have usually been well below the norms because of absorption of funding by salaries and wages. Some observers have also criticized the norms for being too low or not adequately tailored to the specific conditions of particular states and comman. In the future, while cost norms may provide some gener guidance, a better procedre would be to rely on the specific maintenance plans and budgets of each state.  28 (e) M-onitor and control construction quality and costs. The MIS should include analysis of construction costs over time and in different commands, and information relevant to the program for improving construction quality as discussed in Chapter IV. (f) epare financial zirLts. Each ID should prepare an annual irrigation sector financial report, following a standard fornat and including analysis relevant to the points above. The financial report would be used by state governments for annual review of irrigation performance, for establishing improvement objectives, and in preparing and approving the budget and development targets for the forthcoming year. Cost Recovery Policy and PRedu 3.15 The low level and poor collection of water charges has been a persistent and intractable issue affecting nearly all irrigation in India.41 An important technical factor for many commands is poor operations miid maintenance with resultant low impact on farmers' yields, and no impact in the case of tail-enders receiving no water (chapter IV). For customers receiving no service it would be unreasonable to expect a fee, but in most cases some water service is provided and water charges are disproportionately low relative to resultant incremental output (para 3.18). Clearly, however, improvement in water charges is linked to improved maintenance and water management in the case of commands with serious inadequacies in both functions. A second problem is the assessment and collection procedure used. Annual reassessment of water charges is not done and collection often involves complicated procedures and an overstaffed and inefficient government apparatus (para 3.20e). Finally, and probably most important, is the politicization of water charges. Politicians frequently promise farmers not to increase rates or forgiveness of non- payment. This has been a major factor behind both the minimal changes in water charges over time and in the low collection even of assessed rates. 3.16 Irrigation Charges. While GOI's National Water Policy calls for improvement of cost recovery,42 commitment varies within central government and across states. A frequent justiication for low cost recovery is that indirect revenues from irrigated agriculture probably exceed the current cost of operating and maintaining irrigation and, in the case of tubewells, subsidies on electricity.43 This is the case in an agricultural surplus state (Punjab) and may be so, but marginally, in poorer states (Bhatia, 1989). But, India is not in a position where it can afford to forego revenues from productive services that directly increase the incomes of beneficiaries. Further, inadequate charging runs counter to equity concerns. Under good irrigation, farmers gain substantial economic rent from the irrigation service. Water and electricity charges provide a direct way of charging the customers for a service which directly benefits them, whereas indirect taxes also tax non-beneficiaries.44 Irrigation provides farmers with private benefits, and the service should accordingly be paid for. This also conveys legitimacy to farmers' demands for a quality service from irrigation managers and will form the basis for eventual financial autonomy of surface irrigation commands (Chapter V). 41 See Vol. 11, Chapter V for more detailed analysis and discussion of cost recovery. 42 'Water rates should be such as to convey the scarcity value of the resources to the users id to foster the motivation for economy in water use. They should be adequate to cover the annual maintenance and opeation charges and a part of the fixed costs. Efforts should be made to reach this ideal over a period, while ensuring the assured and timely supplies of irrigation water." (National Water Policy Paper, 1987). 43 "Indirect revenues" include agricultural market taxes in various forms (e.g., market yard taxes, "mandi" charges and benefit levies), consumption goods taxes, and taxes on incomes of workers employed as a result of the secondary and teriary employment impact of irrigation. 44 While agricultur market charges may be a proxy for water charges (marketed output is often from irigated holdings), surplus producing rainfed fiamers are alo taxed, which is inequitable.  29 3.17 Charging According to Usage. As elsewhere in the world, "efficiency pricing" of water (equating marginal costs with marginal value product to achieve optimal allocation) has limited scope for application in India. The marginal value of water is high during critical periods and changes unpredictably throughout the season. Nevertheless, both water charges and electricity are currently priced at flat rates and greater efficiency in the use of both, in association with better environmental control, could be achieved by charging variable rates, even where below marginal value. This is evidenced in water-use economy by farmers using diesel pumpsets for irrigation and incurring fuel cost for each liter of water. Practical application for other types of irrigation is difficult. For supply-based systems such as warabandi it has no relevance. For other surface systems, measuring water use by individual farmer is impossible. One adaptation being piloted in Maharashtra45 is to charge volumetrically at the level of a minor, leaving a farmers' group to organize collection of revenues among member families. Such possibilities merit further piloting but would take time to replicate; at present, even assuring water to minors is difficult and minor head structures do not have volumetric measuring devices. Possibilities for charging variable rates are greater for electricity, as electricity metering is technically easier and is common practice in other countries. 3.18 Affordability to Farmers. Farmers can afford water and electricity charges, except in the case of crop failure or where the irrigation service conveys minimal benefits. Present charges are far from excessive; typically less than 2% of the value of crop yields and less than 4% of the net additional rcturns (economic rent) that farmers receive due to irrigation. In most cases, water charges covering the full costs of effective operations and maintenance are affordable. 3.19 In contrast to O&M, the recovery of capital costs may not be possible. While analysis shows that a proportion of capital costs can reasonably be charged for (Vol. II, Chapter V), full recovery requires good incremental production from irrigation, a situation frequently not found, as reflected in economic returns from past irrigation schemes. It should probably be accepted that full recovery of capital investment in most large-scale surface irrigation schemes is unattainable. However, it is important that farmers make some contribution to these investments so that a sense of ownership, participation and responsibility is created. To do so it is recommended that: (i) a nominal charge be imposed on farmers to cover a portion of capital amordzation of the main system to the government outlet; and (ii) farmers make a financial and labor contribution to construction of micronetworks (para 3.2e). Actions to Improve Cost Recovery. 3.20 GOI's longstanding principle of charging for water and electricity at levels at least equal to the cost of operations and maintenance is strongly supported. Clearly, major improvements are needed in its application. GOI should also monitor the performance and reporthig of cost recovery and its impact on the overall financial situation of state irrigation sectors. Water charges should continue to be set at the state level, with increasing application within each state of water charges by individual command. Improvement in cost recovery will be closely linked to improved operations and maintenance of irrigation (Chapter IV). Recommended cost recovery actions are: (a) Set charges at least to cover the costs of effective operations and maintenance in a normal command. "Effective" O&M also covers a nominal share of ID overheads, but excludes excessive state expenditures on wage bills, unproductive expenditures, and rehabilitation costs. Actual O&M costs will vary by command, depending on equipment and service and state of repair. A distinction is needed between costs for rehabilitation, and O&M costs for a "normal" command in reasonable repair. In most situations, water charges could also be set without hardship to farmers at levels that would partly recover capital costs. This would be desirable, though the progressive 45 Majalgaon command under the Maharashtra Water Utilizadon Project (Cr. 1383/Ln. 2308).  30 attainment of full coverage of normal O&M costs would itself represent a major achievement. (b) Introduce water charges specific to an individual comrAand. Higher water charges, including at least part of capital amortization, could be applied following agreement with farmers prior to an investment. The appropriate moment to levy such charges would be at the time of commnissioning a command or completing rehabilitation works or water management improvement. A further step -- retention of water charge revenues at command levei accompanied by a phasing out of government funding for O&M -- could set the stage for eventual financial and administrative autonomy of individual commands. (c) Delink maintenance from cost recovery performance. Adequate financing of maintenance is an essential requirement in itself, regardless of the status of cost recovery. Further, water charges go into general state revenues and are not directly linked to expenditures on O&M. O&M costs are a logical guiideline for cost recovery (the budgetary impact of current expenditures is neutralized), but in situations of poor recovery, maintenance should not be scapegoated. The effect has been inadequate funding for maintenance, based on the excuse of insufficient revenues from water charges to cover these expenditures. (d) Reassess water charges annually. There is no regular mechanism for setting the levels of water charges nor analytical guidance for their assessment. IDs make occasional proposals regarding water charges largely on political grounds. Automatic annual review of water charges should be initiated, based on analysis by a small unit within the irrigation or finance departments, guided by a committee of government representatives (finance, planning, irrigation, and agriculture), and farmers. Clear guidelines for the review would include estimates of O&M expenditures, recurrent expenditure on irrigation, the record of water charges received and collection rates achieved, and an analytical overview of the costs and revenues from irrigation and well electrification, and ti,e implications for the state exchequer of the subsidies involved.46 (e) Reduce Government Involvement in Collection. The heavy government apparatus involved with collection of water charges from individual farmers in many states should be disbanded in favor of using local communities, such as village panchayats or water- user associations (WUAs). In Bihar and West Bengal the costs of colkction are greater than the revenues received (Bhatia, 1989 and Planning Commission, 1991), and this situation is likely in several other states. By contrast, a high collection rate is achieved in Punjab at low administrative cost, using village revenue officials who retain a small percentage of revenues as an incentive. There is also a need to simplify administration of collection. An option for government is to charge each WUA or panchayat an annual fee based on a flat rate per net hectare irrigated. The panchayat or WLTA would be left to make its own decisions and arrangements concerning payments by individual farmers and variable charges by crop or water used. Overall responsibility for collection of water charges from these organizations would be best under the district collector (who has responsibility for other revenues) rather than, as found in some states, under the irrigation department, which should focus on providing the inrigation service. 46 A = forma report shoud be drawn up to assist sta in modelling their analysis; perhaps jointdy by CWC (which has been monitoring cost recovery for some time) and the Ministry of Finance, in collaboration with state governments and financial consultants.  31 (f) Remove the electric power subsidy. The subsidy on electric power for tubewells should be removed.47 Farmers with wells powered by electric motors pay a flat annual rate based on the horsepower of their pump. These rates involve substantial subsidies, for example, 60% on a typical shallow tubewell in Uttar Pradesh (over 75% for more intensive usage). Removal of the subsidy is easily affordable, as farmers get high returns from tubewells, even with diesel pumps, where largely unsubsidized fuel results in a five-fold increase in operating costs, compared with electric pumps. Subsidies for public tubewells (per hectare and, in particular, per unit of water used) are even larger than for shallow electric tubewells. Public tubewells in Uttar Pradesh, for instance, involve complete subsidy for capital costs, maintenance and govemment management, plus, at current estimated collection rates, over 95% of the cost of electricity.48 (g) Implement a variable-cost system for electricity. Previously standard practice in India, a variable-cost system was abandoned due to widespread malpractices. However, the absence of meter-based pricing is causing excessive and indiscriminate use of power, providing minimal incentives to farmers and pump manufacturers to use and manufacture efficient pumps, and is encouraging over-exploitation of groundwater. Full cost meter-based pricing is an important underlying basis for successful groundwater development and exploitation, in many states the principal remaining source of productivity growth through irrigation. It would also have a large beneficial fiscal impact. Consideration of how to achieve this objective is warranted.49 C. Funding Conditions by Central Government 3.21 Central government monitoring of financial management and performance at the state level is also inadequate, starting with limitations imposed by data quality from state governments. Although the ability to conduct strong financial analysis is present in the Ministry of Finance, Reserve Bank of India and Planning Commission, this skill is less present in the Ministry of Water Resources and CWC, both dominated by engineers. Despite their joint concerns, central government fwuds are allocated to irrigation programs with minimal conditionality. Their influence on state expenditures and financial management is scant. Efficiency. Maintenance, and Cost Recovery 3.22 Demonstrated progress in efficiency of expenditure and financial management should be a condition of allocation of Plan funds to the states' irrigation investment programs. GOI will need to improve its capabilities for physical and financial monitoring of irrigation programs in the states. Existing commissions (Planning and Finance) should tighten their review role with enhanced monitoring inputs from MCOWR and CWC and enlisting assistance from the Indian academic community. At minimum, GOI should monitor three measures by each state government to achieve: a) improved financial management; b) full funding of maintenance; and c) progressive 47 See analysis in "Uttar Pradesh Groundwater Development, Issues and Options" (India Agriculture Operations Division, World Bank, February, 1991). 48 The ID provides payments to the state electricity board, but this is merely a transfer payment from one government department to another. The actual subsidy is the full cost of providing power less the revenues received fnom farmers. 49 One way is to charge a high flat rate in the absence of a meter or where a meter is malfunctioning so that this situation would involve payments well in excess of the mte charged even under intensive use with a meter. Village officials could be responsble for collection, retaining a portion of charges as an incentive, and backed by both electricity board and independent monitoring.  32 improvement in cost recovery. Progress would be jointly reviewed by GOI and each state government annually prior to disbursement of Plan allocations for irrigation investment: a) Expenditure Efficiency and Financial Management. Joint review by GOI and each state government of the proposed financial report on the state's irrigation sector (para 3.14f) would foster a continuing process of improvement in financial management and efficiency of expenditures. The financial report would include standard review of physical achievements and expenditures against targets, analytical breakdown of key expenditures and diagnosis and proposed action to tackle identified problems. State governments should also make these reports public to foster greater accountability of IDs to users. b) Maintenance. Inadequate funding of maintenance should be sufficient grounds for not allocating central Plan funds for irrigation development to an individual state. GOI should scrutinize each state's annual maintenance plan and budget (para 3.14d), including field visits by CWC staff, ex-post review of the preceding year's maintenance plan and actual expenditure on maintenance, and review by finance committees of the adequacy of the maintenance plan and budget for the forthcoming year. Good maintenance, including full funding for this purpose, would be expected. For those states encountering budgetary difficulties for maintenance expenditure, despite improved financial management of irrigation and first priority use of state irrigation sector funds for maintenance, GOI could consider allocation of Plan funds for maintenance on the basis of the state's detailed maintenance budget and implementation program. (c) CQst RecovM. Analysis, prepared by the state cost-recovery committee (para 3.20d), of its cost-recovery situation, the total subsidy involved, achievements in the past year and recommendations for the forthcoming year should also be annually monitored by GOI. Desirable features would be quality analysis and wide dissemination of its findings to senior government officials and the political establishment to promote clearer understanding of the fiscal consequences of poor cost recovery and the opportunity to focus on these issues annually, thereby encouraging decisions to implement improvements. The objective would not be to insist on full recovery of O&M costs as a condition for allocation of central Plan funds. While desirable, full O&M cost recovery may not be achievable politically in the short run. Cost-recovery performance should, however, enter into judgment of a state's financial solvency and of its capacity for fursher investment to expand irrigation.  33 CHAPTER IV. TECHNICAL PERFORMANCE A. Productivity 4.1 The productivity of Indian irrigation is below, in some cases half, the levels attained in other large Asian countries and elsewhere. For example, irrigated paddy yields average 2.5t/ha in India, compared to 5.5t/ha in China and 4. It/ha in Indonesia. While care must be used in drawing conclusions from such comparisons (conditions and, in particular, water availability are not the same), they indicate major scope for improvement.50 Irrigation intensity of 1.29 is also low. Additionally, many surface irrigation systems have restricted capacity to enable crop diversification. Without increased yields, cropping intensity, and agricultural diversification, prospects for increased agricultural growth and rural socioeconomic development will be severely limited. Water Management 4.2 Deficient water management is a widespread problem. Some 10-13 million ha of irrigated land (about 55-75% of India's net surface irrigated area) are performing below potential. Very large gains in agricultural productivity can be achieved if water management on these existing irrigation schemes can be improved, and some additional increase in productivity can also be achieved on most of the remaining irrigated area. The most important component for improving productivity of irrigation is better water management, accompanied by upgraded agricultural extension, research, and irrigation technology. Water Distribution Systems 4.3 On most surface commands, water is poorly distributed over area and time. A common shortcoming is that tail-enders are not getting water or are getting insufficient and unreliable water. Conversely, head-enders often get too much water, either because they have no choice (for instance, fields situated near leaking canals or outlets) or by choice, taking water when they can and often more than needed. Also, many farmers have limited knowledge of crop water requirements and detrimentally over-water their crops, even when irrigation water will be available later. Further, the seasonal nature of river flows, compounded by design and water management deficiencies, frequently means that irrigation is only available during and shortly after the monsoon, diminishing possibilities for rabi and summer crops. 4.4 Although the nature and extent of these problems vary widely, they have important impact on agricultural productivity. Where irrigated water is not getting to farmers, production is restricted to rainfed conditions; where land is getting too much water, waterlogging results, reducing yields and causing water table build-up, and eventually salinization. The most pervasive and frequently underestimated situation is where farmers are getting water but in unreliable or poorly timed amounts. Intensification is hobbled; farmers choose low-risk agricultural options: more robust basic cereal crops, hardier but lower yielding traditional varieties, limited fertilizer application, and surer but lower yielding cropping calendars, such as planting after the monsoon begins. Scope for crop diversification is also limited. 50 Some reported irngated yields in official Indian statistics may tend to understate actual yields due to the measurement difficulties and procedures used. Nevertheless, the substantial room for improvement of irrigated agriculture yields in India is confirmed by the large gap noted by Indian resehers between research yields and fanner yields, by the presence throughout India of individual farners with high yields, and by high average yields on commands in India where irrigation has been successful.  34 4.5 For surface irrigation, regional performance is linked to type of irrigation system, agroclimatic and socioeconomic conditions, as well as design, infrastructure and management on individual commands. The main types of schemes in India are: (a) the warabandi system of semi-arid to arid northwest India, where irrigation water is rationed strictly in proportion to farm area and supplied on a predetermined rotational schedule. Farmers decide on crops according to the expected water supply. Infrastructure and operational procedures are relatively uncomplicated; (b) the shgjpani systems of western and parts of central and southern India, where farmers obtain official sanction, usually annually, for proposed cropping patterns and are then entitled to irrigation supplies according to crop needs. Designed at a time when irrigation water was plentiful relative to demand, most shejpali systems are now experiencing difficulties; (c) the localization systems in parts of southern India, focussing on locational control of cropping patterns. Low-lying areas are zoned for "wet" crops (primarily rice and sugarcane), while higher areas are limited to irrigated "dry" crops and more restricted water supplies. Such zoning often breaks down as head-end farmers in "dry" zones take more than their theoretical allocation; and (d) traditional field-to-field irrigation systems used mainly for rice in parts of eastern India and some delta schemes in the south. Continuous irrigation flows are provided, passing from field to field, generally without water courses and field channels. Operating rules have often evolved and been agreed through local tradition, and where water is abundant, yields can be good. However, crop choice and cropping patterns are limited. 4.6 A broad distinction can be made between supply-based or "crop-to-water" systems that distribute water according to predetermined procedures and require the farmer to respond accordingly (such as warabandi) and demand-based or "water-to-crop" systems that attempt to meet crop needs (such as shejpali). In supply-based systems, the role of the irrigation department tends to be simpler than under demand-based systems that require the department to respond to changing farmer needs with more complex and flexible infrastructure and more intensive management. No system is completely demand- or supply-based; rather, the terms capture the tradeoffs between responsiveness and management complexity. Water Management by Region 4.7 Northwest India In the northwest, a good fit has beer achieved between aprolimatic and socioeconomic conditions and irrigation system, and managerient performance is , nerally good to excellent. The average irrigated cereal yield of 3t/ha in Punjab is the highest il. inaia, and probably compares well with yields in other countries if yields are expressed per unit of water used rather than on a per hectare basis. 4.8 Favorable conditions play a large role in the success of the supply-based warabandi system. River flows, fed by Himalayan snow melt from March through October as well as by monsoon rains, provide perennial and fairly reliable flows for the largely run-of-the-river diversions for irrigation. With rainfall limited and irrigation as the most reliable source of water, even in the monsoon, farmers plan on the basis of expected and reliable irrigation supplies rather than rainfall. This limits the stresses that could result from farmer reactions to unreliable rains (para 4.10). Further, due to unfragmented holdings, farmers have flexibility to allocate their water ration within their holdings, enabling some degree of crop choice and good use of available water. Also importantly, the Indo-Gangetic plain aquifer provides ready availability of groundwater for supplementing surface irrigation supplies. Farmer investment in tubewells has grown enormously  35 in the northwest since the 1960s and has provided a major boost to agricultural productivity.51 Access to groundwater compensates for warabandi's limitations in responding to specific water needs and timing of increasingly water-intensive and diverse cropping choices. Higher literacy levels, strong and generally cohesive village community structures, prevalence of land ownership rather than feudal tenancy arrangements, and better availability of transport infrastructure, agricultural support services and inputs also play a role. 4.9 Easter and Centr. Despite higher rainfall, yields, irrigation intensity and possibilities for diversification are poor in eastern and central India, and the difference between rainfed and irrigated yields is often small. Typically, surface irrigation is supplied through iun-of-the-river systems, dependent on rainfall, and usually is only available after the advent of the monsoon, when the problem is often excessive rather than insufficient water. Where well managed, the surface systems could at least provide protective irrigation during dry periods for the kharif crop. But even this limited function encounters major difficulties: farmers' reactions to lack of imely water, design deficiencies, broken infrastructure and poor operational management.52 4.10 Attempts to introduce warabandi in eastern and central India have mostly been unsuccessful. Recent analysis (Berkoff, 199053) illuminates the influence of rainfall on managing irrigation in eastern India and its easier management in the northwest. Rainfall in the Indo- Gangetic plain increases from west to east, and under conditions of higher rainfall (central Uttar Pradesh eastward) all farmers, hoping for sufficient rainfall, plant their entire land to a kharif crop. When a dry interval occurs, however, all farmers need irrigation water on their entire parcels to avoid crop failure. Tremendous stresses arise: head-enders will experience crop failure if they do not take more than their due; tail-enders are deprived of their water and lose their crop. Conflicts arise and breakages by farmers of irrigation structures are common. By contrast, in the northwest, farmers plan on irrigation water as their regular water source, with monsoon showers an expected but unpredictable bonus. Farmers limit water-sensitive crops to perhaps 20-30% of each farm (more, with access to groundwater), with the balance under scratch crops, fodder or fallow. If one takes surface irrigation water out of turn, a specific neighbor's loss (failure of his high-value crop) is greater than the abuser's gain (a marginal increase in a crop that would survive anyway). The system tends to be self-policing and stable, with relatively little farmer interference or damage. 4.11 Peninsular (south and west) India. The performance of surface irrigation schemes in peninsular India is diverse. Management of demand-based systems has often proven difficult under the stresses of different cropping patterns, unpredictable rainfall, more variable topography and soils, and illicit taking of water by head-enders, as well as weak design, inadequate seasonal planning, and poor irrigation management. As in eastern India, breakages of structures, minors and water courses by frustrated farmers is a common problem. Other commands have good yields due to sound design and management and the usual presence of ample water. Higher yields per hectare than in the northwest are obtained in some commands (e.g., Cumbum Valley, Periyar Vaigai and Bhadhra). Where water is ample, average paddy yields of over 6t/ha, and sometimes much higher, are found. However, irrigation effectiveness is lower in terms of yield per unit of water. In peninsular India, rainfed agriculture prevails, but part could also be irrigated if irrigation water was more broadly distributed to support less water-demanding crops. 4.12 Groundwate Iigation throughout Ii. Groundwater irrigation with private tubewells and dugwells generally demonstrates high yields throughout India, mainly due to control by the 51 In Punjab, the number of diesel-powered pumps alone rose from 29,000 in 1968/69 to 263,000 in 1979/B0. 52 Other fatos are coxnplex socioeconomic conditions, prevalence of small and fragmented subsistence holdings often under sharecroping systems, and weak institutions, commercial development and in&frsuc. 53 "Irrigation Management on the Indo-Gangetic Plain," J. Berkoff, 1990 (hIigation Sector Review Background Paper and since published as World Bank Technical Paper No. 129).  36 individual farmer over timing and quantity of water. Conjunctive use of groundwater with surface irrigation has proven effective in making up for the limitations in this regard of surface schemes, providing that surface irrigation supplies are not too far out of tune with crop requirements. As a sole source of irrigation, groundwater has also been effective in areas with good aquifers. Water Management. Research and Extension Issues 4.13 Debate about the merits of different regional irrigation systems has detracted from the more practical consideration of how to improve irrigation performance at the specific command level. The goal is to enhance productivity, using water, a scarce resource, as efficiently and effectively as possible. Each command is different; hence, improving water management requires individual diagnosis and subsequent actions. This procedure, rather than application of a standard blueprint, is a preferable approach. Nevertheless, a number of basic issues need to be addressed, with varying degrees of relevance, depending on the irrigation distribution system and regional characteristics. 4.14 Level of Service and Responsiveness. A basic issue is to decide on the irrigation service's degree of responsiveness to rainfall and specific crop needs. As j iinfall becomes important relative to irrigation, there is a greater need for responsiveness to rainfall. The needs of individual farmers also differ, but design and management to fully achieve this would be very complex, and compromises are demanded in practice. Simplifying design and management to the point where there is little or no response to rainfall or the needs of individual farmers also has its limit; it may serve no one well, unless other features such as groundwater provide the flexibility that a surface scheme does not provide. 4.15 Of necessity there is a trade-off between the concepts of: (a) low response-low management (for instance, supply-based warabandi designed to be simple to operate and have lower needs for control infrastructure and management staffing, but offer little or no responsiveness to rainfall and varying locational and farmer needs); and (b) high response-high management (schemes that provide greater responsiveness based on more control structures and more complicated management). Low-response systems are more manageable in a technical sense, but where local conditions require flexibility, can result in farmer frustration and interference. Where responsiveness to at least rainfall becomes relevant, systems designed for greater flexibility may be necessary. 4.16 Where feasible, it is cheaper and managerially easier to opt for simplicity and options nearer to the low-response system. Yet, a variety of more responsive options exist, if necessary, for local conditions. For example, under the Sardar Sarovar project, irrigation water in Gujarat will be essentially supply-based, but a quick-response, remote-control system will respond to rainfall variations at the block level (500 ha). In the western United States, "modified demand systems" are practiced whereby farmers can purchase water up to certain limits that are adjusted when water is scarce according to pre-agreed rules. This concept might be feasible in India to the level of a minor under farmer management. Experimentation is underway in Maharashtra (Majalgaon command under the Maharashtra Composite Irrigation Im Project, Cr. 1621) to develop a "bulk volumetric distribution system" managed by a farmer group. 4.17 Conjunctive Groundwater Use. Surface irrigation and drainage systems should be planned, designed and operated in relation to possibilities for groundwater usage. Private tubewells or dugwells used conjunctively with surface irrigation can substantially increase agricultural productvity as demonstrated in northern India. A more recent development, drip or sprinkler irrigation usually from groundwater sources, is increasing rapidly in peninsular and northwest India for more intensive horticultural crops. The extent or prospects for conjunctive groundwater irrigation are likely to influence the appropriate design specifications of irrigation and drainage, both at construction and in operational planning of water management. Surface network design should also include consideration of groundwater recharge needs, emphasized by MOWR and MOA as an irrigation strategy objective.  37 4.18 Modernization Optdons. India could benefit from modernization options being applied elsewhere. Relevant options include systems to provide better hydraulic stability, automated water control, and higher water-use efficiency. They do not necessarily require greater managerial sophistication, though they would require more numerous and more sophisticated water control structures (Plusquellec, 1990). A recent case study on Maharashtra (Baudelaire, 1990) estimates that a more precise and responsive system with incremental investment costs of 10-15% might enable substantial gains in productivity. Features that would reduce the need for incremental management were also considered possible. Little experience is actually available in India on such options. Nevertheless, examples of successful experimentation and practical application elsewhere, and the evident need to improve water management in India, point to the need to do further applied research and piloting for innovative water management systems. 4.19 Irrigation Research and Networking. Applied and operational research on irrigation and drainage is lagging in India. This partly explaiins the 1,mited application of new technologies in widespread practice elsewhere. More applied research and pilot efforts are needed on mechanisms to improve water management, maintenance, drainage (including salinity and sodicity control), control of siltation and the environment. Continued adaptive rescarch is needed on construction matters (materials research and testing, canal linings and seepage). In addition to technical research, more piloting efforts and case studies are needed to support social and organizational innovations in irrigation (for instance, in development of water-user associations and involvement of women). 4.20 Many institutions are involved with irrigation research; state engineering research institutions, state Water and Land Management Institutes, national institutes such as the National Institute of Hydrology, a variety of project-specific research programs, related research done by agricultural universities, and central institutions such as the CWC, CGWB, and Central Board for irrigation and Power (CBIP). Given India's diversity and the need for locally responsive research, this array is probably necessary. Efforts, however, are disparate and sometimes duplicative, and require better linkages to specific operational problems. 4.21 Recognised needs are to increase the relevance and technical quality of research and, as important, the networking of research findings both between states and institutions within India and through linkages with research, technology and irrigation management developments in other countries. To this effect, MOWR is considering stepped up funding of irrigation research and better mechanisms with supportive funding for networking of research and technology developments. Useful initiatives by CBEP and CWC to disseminate research results across states and encourage inter-state linkages, need broader and more systematic application. Networking would emphasize innovatory ideas for irrigation technology, irigation management, financial management and institutional development. This would include workshops to share and integrate ideas from academics and useful experience from other countries.54 Study tours, sabbaticals for research or university study, and operational secondments should also be funded. 4.22 Stepped up funding of research and related training and dissemination would need to be accompanied by creating institutional mechanisms for the sharing of ideas and for programming and funding research and technology transfer priorities. This should be coordinated with the activities of the Indian Council of Agricultural Research and the state agricultural universities. Private sector research and technology transfer and NGO initiatives should also be encouraged. For instance, the rapid development of drip and sprinkler irrigation has been stimulated by research, demonstrations and extension by private equipment manufacturers. MOWR has recently S4 Workshops would cover both specific technology issues and the shanng of expenence in a variety of management aspects; for instance, construction supervision and quality control measures being applied in the Sardar Sarovar Project, West Bengal's experience in tansfer of deep tubewell management to panchayats and experience with private and public tubewells in other states and countries, successful examples of cost accounting and management information systems, etc.  38 had informal contacts with the International Program for Technology Research in Irrigation and Drainage (1PIRID), sponsored by the World Bank, UNDP and the International Commission on Irrigadion and Drainage. IPTRID, which provides technical assistance to interested countries for establishing research programs, could provide a further opportunity to expand linkages with international research and technology development 4.23 Improving Agricultural Extension. To realize the full benefits of irrigation investment, agricultural extension and other support services need considerable improvement. Field observations indicate that on most commands extension impact is still weak; many farmers are insufficiently aware of optimal water use and fertilizer application, trace element requirements, planting density and time of planting. On commands where extension services are high-quality, better agricultural practices are observable. Knowledge of optimal water requirements is particularly weak, as neither irrigation nor agricultural extension services provide such advice. Farmers tend to overwater crops if they have access, with detrimental yield and environmental consequences, partly due to risk-avoidance but also to limited knowledge of best practice. 4.24 Ongoing Ministry of Agriculture and state agriculture department efforts to strengthen agricultural extension services, not in numbers but in terms of the quality of their advice and relevance to the specific needs of the irrigation command concemed, should be continued. This should include training of extension staff in optimal water application for crops for subsequent dissemination to farmers. MOWR has also recommended enhanced use of on-farm demonstrations in extension, creation of command level farmer training centers, greater use of the media (radio, television and newspapers) which has proven an effective supplementary source for transmitting extension messages, increased use of available state soil testing services to refine nutrient use in command areas, and the conducting of soil profile surveys to help define water delivery systems and drainage needs. All such initiatives will require improved coordination between MOA and MOWR at the center and between agriculture and irrigation departments at state levels. Actions to mrove Water Management 4.25 The first priority is to implement widespread and fairly rapid improvements in water management. Given the vast irrigated area, improvements must be low-cost. This requires emphasizing water operation planning and daily management, supplemented by low-cost infrastructural improvements to facilitate better control of water. Recommended actions are to: (a) Implement improved water operation plans and daily manage.ment (b) Implement low-cost infrastructural improvements to facilitate better control of water. (c) Selectively opt for higher-cost infrastructural improvements. To the extent that funds are available, more costly infrastructure for more intensive water management to achieve greater responsiveness to rainfall and farmer conditions would improve productivity. Limited experience to date in India with such irrigation systems would require stepped up applied research and piloting. In the short term, such a program would be limited in extent both by the need for piloting and by funding limitations. (d) Step-up applied research. networking and piloting of irrigation technology. (e) Emphasize agricultural extension on optimal water usage. General agricultural extension also requires improvement on many commands. 4.26 The National Water Management Project. The National Water Management Project (NWM,1P, Cr. 1770), is piloting the implementation of some of these priorities. The project aims at helping participating states to manage existing irrigation schemes better, specifically by improving the capacity to plan, implement and monitor operations on selected commands and by funding associated low-cost infrastructural improvements. The project uses a command-by-command  39 approach, involving a process of diagnosis, prognosis and remedial action. At scheme level it involves the following steps. Diagnostic. A command-specific diagnosis is made of current operations, the water needs of farmers, and deficiencies relative to these needs. Qpration PaPn. A prognostic program for improving scheme management and implementing low-cost infrastructural responses to identified needs. The central element is an "operation plan" on how the scheme is to operate through the year and to adapt to contingencies such as droughts or floods. Infrastructural Improvements. An average upper limit of Rs 4000/ha (about US$ 200) is budgeted for infrastructural improvements, generally confined to repairing or adding control structures in the conveyance system and providing communications and monitoring equipment. Managerial Improvements. Adaptations in scheme management generally require staff training, adjustme-t of staff activities, mechanisms for better supervision and management information, and a .nonitoring and evaluation system. 4.27 The NWMP is presently confined to Kamataka, Tamil Nadu and Andhra Pradesh, but additional states have recently joined: Uttar Pradesh, Bihar, Madhya Pradesh, Orissa and Kerala. Implementation so far has been slow and performance mixed; generally encouraging in Karnataka and Tamil Nadu, but slow in Andhra Pradesh and in the initial planning stage in most other states. The slow initial progress partly reflects the low priority traditionally given to management improvements relative to construction of infrastructure. The project has, however, raised consciousness of the fact that significant productivity improvements are possible in existing schemes at limited cost.55 Under the Eighth Plan, a large centrally sponsored irrigation management program, to be known as the "Assured Irrigation Scheme," is being considered. NWMP would eventually involve 0.6 million ha, compared with 10 million ha targeted under the Assured Irrigation Scheme. A major issue in implementation is how the program is to expand rapidly in view of substantial planning, training and management needs when such skills among govemment staff and local and foreign consultants are in short supply. MOWR is establishing a national committee to work out how these constraints can be resolved so that the program can be accelerated and expanded while assuring quality impact. B. Sustainability 4.28 The enormous investment (nearly 10% of total Plan outlays since 1950) made in India's irrigation infrastructure must be sustained and its ability to perform enhanced. Irrigation works must be designed, constructed, operated and maintained to ensure continued good performance over time. In the absence of a complete inventory, field observations by government observers and Bank staff conclude that in nearly all states, with the possible exception of Punjab and a minority of surface irrigation commands elsewhere, irrigation and drainage infrastructure is deficient and deteriorating. The principal reasons are poor inital design, poor quality of construction, and inadequate maintenance. If Ehns situation persists, irrigation infrastructure will continue to decline, resulting in reduced ability to perform and either a reduction in surface irnigated 55 Monitoring in terms not only of impact on water distribution (timing, reliability and area irrigated by season) but also in terms of yields and crops grown needs to be stepped up. Nevertheless, initial results from some commands are encouraging. Bhadra command in Kanataka, which has been under the NWMP since 1987/88, won the 1991 National Productivity Council Award for high production under irrigated agriculture (an average of 6 t/ha of paddty in each crop season) and with irrigated area on the command increasing by an average of 15% (ranging by distributary from 1% to 43%). (See Volume II Annex 6.3).  40 area or a prohibitively expensive rehabilitation program. Already, rehabilitation needs are considerable. 4.29 Sustainability also has environ.. .vntal, financial and institutional dimensions. Environmental deterioration is occurring on a number of commands due to lack of drainage to counteract waterlogging and salinization; other adverse environmental impacts are also occurring. Activities such as maintenance must 'e affordable, implying efficient use of funds, adequate generation of revenues from irrigation and the maximum handling of simple maintenance activities by beneficiaries. Institutional adaptation is also required to provide functionally specialized divisions focussed on design, construction and maintenance, and capabilities in environmental assessment. Additionally, farrners and local communities need to be incorporated into decision- making and management so that their influence is contributory rather than passive (lack of participation in operations and maintenance) or disruptive (illicit taking of water and breakages of structures). Design 4.30 Most design problems stem from inadequate data and unrealistic assumptions about water availability and irrigation efficiency. More complete basic data on hydrology, rainfall, soil characteristics and cropping pattems are needed. Then, agricultural objectives can be more carefully considered: the water needs for existing cropping patterns and likely future diversification; complementarity among rainfall, surface water and groundwater; and planning for conjunctive groundwater usage and, as necessary, groundwater recharge. Drainage must also be built into the project design, even where investment in drainage is deferred. Finally, socioeconomic conditions must be analyzed in consultation with the local community and features responsive to local circumstances incorporated (farn size and fragmentation, tenancy systems, literacy and poverty levels, social structures and the needs of women and minorities). 4.31 More realism concerning the availability of water and feasible efficiency of water usage is in order. Concern about ensuring an adequate benefivcost ratio, exacerbated by political concerns to maximize planned irrigated areas, adds pressures to overextend proposed command areas and use unrealistic design assumptions. Irrigation efficiency56 in India has often been assumed at 60%, whereas a worldwide sample of irrigation commands indicates 37-40% efficiency in regions of low rainfall (below 1000 mm) under reasonably good management, and in higher rainfall zones, an average of 23% (Bos and Wolters, 1989). Most irrigation commands in InCia probably have an irrigation efficiency of 20-35%. If assumed efficiency is 60% and actual efficiency is 30%, actual water availability will be half the assumption at design. Another common deficiency is that potential irrigable area is often based on a standard 75% probability level for water availability. This has no necessary relationship with what is optimal for the command (only simulation analysis can determine this) and also often results in overestimation of potentially irrigabie area. 4.32 The above practices substantir'!y explain why many surface irrigation schemes cannot perform as hoped for. Design deficiencies are also partly behind the "gap' noted in some government commentaries between assumed created irrigation potential and actual irrigated area. Many schemes cannot deliver water in the amount as planned, and for these situations the command area targeted for coverage should be reduced to cater to these realities. Construction Oualy 4.33 Construction quality varies widely. While there are some examples of good quality construction (amongst them, Bhakra-Nangal, Beas Darn and Beas-Satluj link projects, and 56 The percentage of water actually used in a beneficial way for crop growth relive to water delivered at the scheme intake. Nonnal losses from seepage, evaporation, evapotranspiration and spillage occur in every irrigation system, but are increased where a scheme is poorly designed, maintained and operated and by poor field application.  41 currently the Sardar Sarovar r.ojectS7), inferior work is more common. Construction quality has deteriorated sharply, as evidenced by recent field inspections and by the frequent observation that works most in need of renovation or reconstruction were constructed in the past two decades. Poor construction is particularly prevalent with routine works such as minor, branch and major canals, including control structures and (where present) linings. Construction quality is also deficient for a number of dams, posing serious potential risk to downstream populations unless rectified. (GOI and state initiatives to address this problem will be assisted under the Dam Safety Project58.) In some cases, renovation of deteriorating works is required before projects are commissioned. By contrast, infrastructure construct.ed 30-100 years ago is often still sound, even when abutting deteriorating modem day features. 4.34 Attaining acceptable construction quality involves little additional expenditure. If the original design is good, it is not a question of additional materials but of a quality control process to ensure construction according to the specifications. Not achieving good quality of construction is, by contrast, highly expensive. Failure requires rehabilitation or reconstruction and for dams, presents obvious risks. Tackling the problem will require more rigorous management procedures and site supervision and increased staff training in some functions. This is certainly within the capabilities of the Indian irrigation engineering cadre, as evidenced in most states by examples of technically complex dams, canals and control structures built to good standard. 4.35 Financial factors and local pressure to take shortcuts or ignore poor quality work are the overriding negative influences on construction. During construction, large illicit profits can be made by using substandard materials, skimping on the thickness of linings, and labor-saving shortcuts in curing cement. Technical and procedural improvements in irrigation departments will have only limited impact on construction quality if such influences on staff are not tackled. Creating better awareness at the state level of the consequences of these pressures is a necessary first step. Firm political will to do something will be necessary, accompanied by top-down leadership from the irrigation department hierarchy, with strong political backing and support from senior authorities. With this support, technical improvements by irrigation departments in construction design, contracting and (most importantly) in quality control during construction can substantially improve construction quality (Price, Malhotra and Fauss, 199059). Actions to Improve Construction Quality (a) Apply rigous quality control, during all phases of construction, with meaningful site inspections on at least a daily basis by the H) staff responsible for works supervision. Good internal quality control is the most essential need. Authority and accountability of 57 The successful transition in the Sardar Sarovar Project (Cr. 1552/Ln. 2497), from initially problematic quality control to good quality construction now, is illustrative of the measures that need to be taken to improve construction quality. These mekures included careful selection of staff and consultants to provide full competence in all required specialties, strict monitoring of the quality of cement, insistence on use of proper equipment by contactors, rigorous on-site testing and inspection, enforcement of contractor adherence to specifications in the contract, and dismissal of poorly performing contractors. Critical to success was strong back-up to on-site staff by the Gujarat state irrigation hierarchy and the political authorities. Professionalism was both insisted upon and rewarded and, in particular, quality control measures were taken seriously by state authorities, enabling staff and contractors to emphasize and implement quality works. 58 The project (Cr. 2241/Ln. 3325, FY91) will assist participating states and the Central Water Commission to improve institutional capabilities to inspect and evaluate existing dams on a regular basis, upgrade the safety of distressed dams identified under the program, and expand the national flood forecasting network Four states will initially participate (Madhya Pradesh, Rajasthan, Tamil Nadu and Orissa). Other states may join in a follow-up operation. 59 "Improving Construction Quality of India Irrigation Civil Works," Price, Malhotra and Fauss (Irrigation Sector Review Background Paper, 1990).  42 construction supervisors must be clearly designated and an "OK" card system used for all key actions and materials used. As necessary, equipment and laboratory facilities for quality testing and additional staff training should be provided. Site supervisors should be provided with supportive advisory visits and inspections by more senior staff, and rewards and sanctions applied against staff performance. (b) Improve construction design. More complete compilation and analysis of basic data; care in the evaluation of site conditions and practical designs responsive to local conditions; clarity and completeness of plans and specificetions; and the packaging of contracts to facilitate high-quality work are all required. (c) Tighten contracting procedures. Bid packaging and selection practices should include the criteria of technical competence and financial soundness (small firms often do not have the equipment and resources to do good work). (d) Dismissal of poor contractors. Poor quality should be subject to immediate suspension of payment and rectification and, if necessary, dismissal for non-compliance. (e) Increase public involvement and accountability. An independent process to audit quality should be introduced, involving consultants or an independent agency. Local communities and farmers should also be involved at each site. For the bulk of construction, local communities, which have direct interest, can assist ID supervisors by monitoring and applying community pressure on quality adherence by contractors. (f) Formalize quality control procedures. Additionally, MOWR has recommended preparation of manuals providing guidance to ID construction divisions and involved engineers on procedures to be used in all aspects of construction quality control; in particular as concerns on-site supervision and quality inspection/accountability procedures, but alsc including guidance on appropriate features in bid preparation, contracting procedures and measures for contract enforcement. Maintenanc 4.36 Signs of poor maintenance -- silt deposition, weed infestation, broken linings, malfunctioning and damaged structures -- are visible in most surface schemes, and many drains are no longer discemible. Pumping equipment in many state tubewells is also out of operation. Poor maintenance makes it impossible for irrigation and drainage infrastructure to control flows and deliver and dispose of scheduled water supplies, and results ultimately in reduction of irrigated area or in the costly need for rehabilitation. 4.37 Central Government and most states have expressed long-standing concern about irrigation maintenance.60 Yet these concerns have rarely been translated into effective planning, budgeting and action in the field. Tackling maintenance requires confrontation with politically difficult realities and executing u-npopular actions, including halting unnecessary staff growth and addressing corrupt contracting practices. The latter is difficult to control as actions are dispersed across each state's irrigation and drainage network. Also, the results of poor or no maintenance are not readily visible in the short tern. For politicians, the short-term benefits of maintenance are less tangible than the politically more attractive option of new construction. As a result, improving maintenance has remained a low priority for state govermments. 4.38 Farmer Participation. Farmers are usually treated as passive recipients of irrigation and have not been adequately consulted in design, construction, operations, and maintenance. 60 "Maintenance of existing irigation schemes is not getting the attention of states as required" (Public Accounts Committee, 1983); and "One of the main reasons for underutilization of irrigation potential was that the maintenance of irfigation and drainage systems was neglected" (GOI, 1990).  43 Consultation with local communities has improved in many states, with positive results, but much more can be done. Farmer participation is not a substitute for quality design and operation of irrigation schemes. It can, however, substantially improve the quality and acceptability of irrigation systems, as infrastructure and water management stand a better chance of being tailored to community and agricultural needs. Operational plans need to be fully discussed, understood and agreed on by farmers on an annual basis, and especially whenever significant changes need to be made. Farmer involvement creates an important sense of community participation, responsibility and ownership. ilicit taking of water and breakages by farmers or sub-groups are less likely where the community as a whole is involved in decisions and system management and implementation. 4.39 Maintenance of Micronetworks. Good maintenance of micronetworks (watercourses, field channels and field drains) is also important. The vastness of such networks imposes financial, staffing and administrative constraints on the irrigation departments' capacity. Although commonly recognized, in some states, government maintenance operations are increasingly being done on micronetworks. This is not financially and administratively tenable over a large area or over the long term. Instead, efforts should be made to encourage and assist farmers to do the maintenance themselves. 4.40 Maintenance of micronetworks by farmers essentially requires a functioning and well managed main system to the government outlet so that farmers get a good irrigation service; only then will it be worthwhile for them to maintain the system at their level. Secondly, farmers must be encouraged to own their micronetworks: ideally construction by farmers, or at least a labor and fmancial contribution to construction. Thirdly, farmers need a grass-roots institudon such as a water-user associadon (WUA) to organize mutual self-help in handling construcdon and O&M (Chapter V). From govemment, farmers need technical assistance and training in construction and maintenance techniques and help in formation of WUAs. Provision of credit (and perhaps matching grants) for investments or improvements would also be needed. Actions to Improv Maintenance 4.41 As with construction, maintenance will only be improved if the political will is there to tackle the problem and provide irrigation departments with the backing and financial support required. State and central governments must also get more involved in monitoring maintenance performance by IDs. Technical aspects are well understood; the missing elements are inadequate prioritization and loose planning, budgeting, supervision and monitoring. (a) Make maintenance first priority, including attendon to management, staffing and budgeting (first call on state irrigation expenditure, Chapter Ell); (b) Tighten IDs' planning. budgeting and performance monitoring (using the proposed annual maintenance plan, budget and performance review, to be monitored by state and central governments, Chapter D); (c) Establish or strengthen a special maintenance division in each ID. Allocate high-quality managerial and technical staff to this division, including financial analysis and monitoring capabilities; and (d) Sup=ort micronetwork maintenance by farmers. Extension and credit support is required. A decision is required, followed by staff reallocation, on which agency (irrigation or agriculture departments or CADAs) will provide the extension support (Chapter V).  44 Environmental Issues and Trends 4.42 The environmental and resettlement impacts of irrigation are widely debated, but due to political perspectives, and lack of data, discussion has often been polarized. A balanced yet more operationally responsive approach to environmental and resettlement needs is required.61 4.43 Irrigation's Impact. Irrigation has probably had a net positive impact on the environment, largely because of its role in increasing production of food and biomass (fodder and fuel) for a growing population, thus reducing pressures on rainfed agricultural and forested land. Erosion on marginal soils, overgrazing and deforestation would have been far more acute without irrigation. Other positive impacts include socioeconomic, nutrition and health benefits from higher incomes, better and more secure food production, and village water supply. Substantial hydroelectric power is also a product of many storage-based irrigation schemes, reducing reliance on environmentally inferior thermal power stations. 4.44 Yet there are negative aspects of irrigation development, among the most evident being waterlogging and salinization where drainage is inadequate or water distribution poor, population displacement and loss of land (including forest) for canal and drainage infrastructure or inundation for reservoirs, and water-borne disease risks. Other deleterious impacts may also occur (Blinkhorn and Rees, 1990). 4.45 Waterlogging and Salii ization. Situations where irrigation has caused groundwater build- up and induced waterlogging and salinization are now found on 3% of existing command area (Smedema, 1990) due to inadequate drainage. The situation is particularly serious in parts of the northwest,62 and some localities can no longer be cultivated. Lack of drainage is also responsible for uncultivable saline and sodic soils ("usar lands") in T'.tar Pradesh.63 The much more widespread problem is waterlogging of both irrigated and rainfed lands during the monsoon exacerbated by lack of drainage. Installation of drainage, neglected in the past, is key to solving these problems. Waterlogging can also be reduced through improved water distribution from the irrigation system and conjunctive use of groundwater. 4.46 Land Losse and Population Displacement. Lands lost to canal and drainage infrastructure under run-of-the-river type surface schemes typically represent 2-5% of the irrigated command area created. With schemes involving reservoirs, a further 3-8% of land is lost. The area submerged depends on the dam site and the size of the project, with better topography and larger reservoirs tending to involve smaller ratios of loss to irrigation potential created. For medium projects, submergence ratios are usually greater than 6%. By contrast, the very large reservoir being created under the Sardar Sarovar project, which has ideal topography as the dam site blocks a deep valley, will involve land inundation of only 2% of created command area. The extreme is found with some small tank irrigation schemes in southern India where a unit of land may be submerged for 61 Mote detailed discussion of environmental and resettlement issues is in Vol. n, Chapter VI, and in the following Irrigation Sector Review Background Papers: "Environmental Issues Related to the Irrigation Sector," T. Blinkhorn and C. Rees (1990); "Land Drainage and Reclamation," L.K. Smedema (1990); and "Land Acquisition, Resettlement and Rehabilitation," W. Partridge and A. Salam (1990). 62 In parts of Punjab and Haryana and localities in Rajasthan and Gujarat, water tables have risen by 25-30 cms/year since the late 19th Century (Vol. H, Figure 6.1). No deleterious yield impact generally occurs until the water table is near the surface, a problem that must now be faced in some areas through investment in drainage (Volume H. paras 6.5 to 6.11). 63 Where present, irrigation will have exacerbated development of usar lands, but lack of drainage is the principal cause in Uttar Pradesh.  45 each unit irrigated. For India as a whole, an average of six families are displaced per 100 families provided with surface irrigation, and the livelihoods of displaced populations must be protected.64 4.47 Forest Loss. The Forest Conservation Act of 1980 prohibits conversion of designated forest land65 to non-forest use, which frequently results in blockage of storage-based irrigation projects and complicates resettlement efforts. Expansion of irrigation and other uses of water will inevitably require submergence of land for reservoirs, in many cases, designated forest lands, thus running counter to forest policy. Deforestation concerns are legitimate and their extent underappreciated. India's remaining forest resources are rapidly being depleted by tree felling for local household needs, overgrazing, inadequate control of commercial timber interests, and agricultural encroachment by poor farmers.66 According to government estimates, up to 1.5 million ha of forest are lost annually. However, total land of all types, including forest, lost annually to reservoir inundation is estimated at less than 50,000 ha (less than 4% of annual deforestation). 4.48 Siltation. Upstream soil erosion, mainly on rainfed lands, has caused much higher than anticipated siltation of reservoirs and canals. On some irrigation schemes MOWR has commented that the problem is exacerbated by inadequate design features for sediment control. The most fundamental need, for the sustainability of both rainfed and irrigated agriculture, is for improved soil and water conservation practices on rainfed lands, pastures and forest cover.67 Specific to the irrigation systems, MOWR intends to support stepped-up monitoring and research and remedial actions as practicable for sediment-control, starting with the most affected reservoirs and canal systems (Vol II para 6.13). 4.49 Water-Borne Diseases. Water-borne diseases, such as malaria, schistosomiasis, typhoid, diarrhea and filariasis are presumed to be a risk of irrigation, although limited information on impact is available.68 Studies of the few areas irn India affected by schistosomiasis show that the disease is not spreading, but the situation must continue to be monitored. Data on other diseases are diffuse and often inconclusive, though knowledge of vector life-cycles indicates that risks are present More monitoring and research of such risks and appropriate precautions against them are 64 Based on analysis of the impact on population resettlement needs of 11 World Bank-assisted projects financed between 1978 and 1988 (Volume U, Table 6.2). 65 Lands classified as forests. Depletion of tree cover has, however, resulted in deforestation of much of the designated forest land. 66 Discussions of deforestation and ecological degradation issues in India are contained in the following World Bank reports: Wasteland Development Review (1988), Review of Rainfed Agriculture and Watershed Development (1988), and Agricultural Technology Review (1989). A Review of the Forestry Sector is planned for 1992. 67 The critical need for improved soil and water conservation has frequently been commented upon by the Ministries of Agriculture, Environment and Forests, Water Resources, the Planning Commission, state governments and the World Bank. Specific initiatives supported by the Bank have included research and piloting of vegetative on- farm barriers such as vetiver grass, promotion of contour ploughing, the forestry projects portfolio, research and promotional features under the Research and Extension Projects, four Watershed Development Projects (Pilot Rainfed, Himalayan Watershed, Hills and Plains), sector reports related to these issues (e.g., Watersheds, Wastelands, Agricultxal Technology) and the current initiative to prepare a Narmada Basin Development ProjecL This would be focussed on agricultural and forestation measures to preserve rainfed soils and vegetative cover, increase soil absorption of water and increase rainfed agricultural productivity. 68 Positive health impacts of irrigation have also been commented on and most observers consider that these far exceed known negative health impacts; diets are improved, water is made available for washing and hygiene, and groundwater is replenisLed enabling extraction for safe drinking. Nevertheless, known or suspected disease risks from irigation must clearly be tackled.  46 needed. Risks can be substantially reduced by removing sources of stagnant or slow-moving water, frequently controllable by proper maintenance of drains and canals and efficient water management. 4.50 Other Environmental Impacts. Among other environmental problems cited, each meriting closer monitoring, are: chemical pollutants, such as nitrates from fertilizer, pesticides and industrial waste carried by water, river flows affected by irrigation (in particular, the situation in the Ganges basin needs monitoring); and loss or change of flora and fauna habitat. Tackling Environmental and Resettlement Issues 4.51 Innovative steps conc rning environmental policy and successful piloting for resettlement and rehabilitation of displaced communities have contributed to significant, though piecemeal, progress by some states. For both issues, national policy frameworks, prepared in consultation with state governments, are required to ensure consistency of approach and widespread application. Resettlement and rehabilitation policy also needs a national legal framework. Environmental impacts have no necessary linkages to state boundaries, and such broad considerations cannot be tackled on a project or even specific-state basis. River basin planning and implementation (Chapter II), including environmental concerns, will be an important basis for environmentally sensitive and sustainable irrigation. Recommended actions below, while facilitated by national policies, do not need to wait for national policy papers to be finalized. 4.52 Environmental Policy. Since 1985, the Department of Environment of the Ministry of Environment and Forests has mandated an environmental impact assessment for all irrigation projects presented for GOI financial assistance.69 While greater environmental awareness, skills and knowledge are achieved in the process, several drawbacks need attention: First, overuse of the environmental clearance to block projects with potentially negative impacts has inadvertendy led to circumvention by political pressures and approval of environmentally inappropriate projects. Earlier assessment of environmental impact, with a view to incorporating design features to minimize environmental damage, would be a better approach. If satisfactory solutions cannot be found, project blockage may still be appropriate, but this decision needs more measured consideration of full project costs and benefits, including the environment, rather than the single yardstick of environmental impact. Despite environmental assessment, project implementation has often fallen short of adequate handling of environmental concerns and objectives. Follow-up environmental monitoring and remedial actions are also required. Finally, the environmental assessment procedure, while useful for project-level decisions and implementation, is too limited, and requires a broader context. Environmental understanding is beset by a lack of data and insufficient skills in measurement and analysis. The institutional apparatus for 'iandling environmental analysis, making decisions, and funding and implementing environmental research is still limited, and river basin planning has hardly been attempted. 4.53 Actions to improve environmental assessment and impact. (a) Prepare a national policy for environmental management in irrigation to be done joindy by MOWR and Ministry of Environment in coordination with state governments. State governments should subsequendy ratify this. (b) Address specific environmental concerns through project investment. The largest future need will be for drainage investment, already a large component in several World Bank-financed projects (e.g., Second Punjab Irrigation and Drainage Project, Cr 3144/Ln 2076, and Narmada Water Delivery and Drainage Project Cr 1553). Specific measures dependent on project location, type and local environmental issues will also be required (for instance, the Upper Krishna Phase II Project, Cr 2010/Ln 3050, 69 The World Bank also has an Envilrnmental Assesmnent procedure conducted during project prepartion.  47 includes pilot actions to control against malaria). For all projects, any necessary project- related environmental monitoring or applied research should be included.70 (c) Conduct environmental impact assessment as early as possible in project identification and preparation, make it integral to this process, and use it to influence project selection and design; (d) Expand environmental monitoring and applied research by the states. supplemented by nationally implemented work. Increasingly, monitoring should be handled on a river- basin level. Central guidance could be provided by MOWR and its implementation agencies in coordination with the Ministry of Environment; (e) Improve institutional capabilities to address environmental concerns. At the state level in particular, such capability is still rudimentary. Adaptation and strengthening of existing institttions would be preferable to creating new ones. Staff need to be trained, institutional roles and interrelationships defined, and financial resources provided for data collection, analysis, research and monitoring; and (f) Recognize and use pricing mechanisms for environmental management. Pricing of electricity for wells is a serious issue. Market signals to private users offer the only mechanism that can be expected to curb overexploitation, a problem that has already had deleterious effects on the water table in numerous areas.71 Although administratively difficult, meter-based full pricing of power would have positive environmental, as well as fiscal, impact; 4.54 Policy on Land Loss. The issue of land loss and its impact on productivity, the environment and local populations, while important in assessing irrigation proposals, should be recast in the broader context of optimal land use and the costs and benefits associated with alternative land uses. The direct and indirect costs of land (and forest) losses have to be factored into the rate of return calculations used in assessing project proposals. Costs and benefits may vary considerably, depending upon type of project, the ratio of lands lost to lands brought under irrigation, population density, the expected productivity of irrigation investment, and the economic, ecological and social value of the lands that would be lost. The question is more complex if value is difficult to quantify, as in the case of the ecological value of well forested lands, lands holding endangered flora and fauna, or lands of historical or spiritual significance. However, all such assessments should be made on as full an analytical base as possible, taking into account both costs and benefits, including full presentation to decision-makers and the general public of non- quantifiable factors. 4.55 Eoae1Yy Polcky. Several modifications to forestry policy are recommended. First, closer examination of tree cover in designated forest land is needed. Much designated forest has been substantially degraded and this should influence the evaluation of ecological impact or potential reservoir site. Second, afforestation as an associated investment with irrigation development should be considered. Under the Sardar Sarovar project, for instance, for every hectare inundated, two hectares of land will be reforested. 70 Environmental monitoring needs will vary depending on the command and its agroecology. A common need recommended by the Ministies of Water Resources and Agriculture is for soil profile surveys to detemine the degree and extent of waterlogging and drainage needs, or, conversely, needs for groundwater recharge. Groundwater levels need monitoring more generally to determine trends of diminishing or mounting groundwater levels. 71 Western Uttar Pradesh offers an example of declining groundwater tables (by 30 to 50 cms per annum in some areas). Some scope for controlling overexploitation of groundwater is available from administrative measures. As at present, credit for investments in wells can be restricted in areas where groundwater tables are declining. But this will not affect investm3nts witnout credit unless such investments can be prohibited. Such policies are likely to be only partially effective, especially over the long run.  48 4.56 Resettlement and Rehabilitation (R&R) Policy. Resettlement and rehabilitation of communities displaced by irrigation has rarely been satisfactory. For those displaced due to development of canals, drains and inundation for reservoirs, alternative government land has often not been available. While cash compensation has usually been provided to land owners, it has generally been inadequate to purchase new land. Further, tenant farmers and landless families often receive no compensation, and physical infrastructure and services for rehabilitation have seldom been provided. This situation is unacceptable. It has also compromised the general public's trust and is a major factor behind much of the controversy over current irrigation development. Unless this issue is resolved, it will only worsen due to population pressure. The fact remains that the creation of water storage is critical to future development of water resources in most regions. However, the welfare of all displaced persons must be protected, and ideally they should become net beneficiaries of irrigation development. 4.57 The need for change is now generally recognized. At field level, some advances have been made, notably in Gujarat under the Sardar Sarovar project. Features include: designation of a minimum land holding for resettlement (2 ha of irrigated land per resettled family member, including landless families); adequate funds for purchase of land; assistance in R&R through Land Purchase Committees; and active support from government staff and NGOs. Public infrastructure (roads, electrification, drinking water, schools, etc.) is also being created. Detailed planning and establishment of institutional capabilities have provided the necessary backing for these efforts. While initially slow and requiring adjustments, implementation in Gujarat provides valuable lessons for replication elsewhere.72 4.58 Encouraging examples are few, and R&R still lacks an underpinning policy and legal framework. Overall, funding is limited, planning rudimentary and inistitutional arrangements for implementing R&R need to be developed. Nevertheless, welcome initiatives are being taken. In 1989, a Draft National Policy on Development Resettlement was prepared jointly by NGOs and the GOI Commissioner for Scheduled Castes and Scheduled Tribes, followed by a draft proposal on revised national R&R policy which GOI recently discussed with the Bank. Several states (Maharashtra, Karnataka and Madhya Pradesh) have passed legislation defining policies for resettlement or have commenced deliberations on future R&R policy. Such initiatives, and practical experience gained under Sardar Sarovar, provide a basis for a more comprehensive approach to R&R and its effective implementation in the field. To provide a framework for guidance for state implementation of good R&R programs, MOWR has established a committee on R&R through its National Water Board, also involving other concerned ministries. In consultation with the states, the committee will produce a National R&R Policy Paper (para 4.59). MOWR also intends to encourage each state to undertake a case study of one of its R&R programs and to facilitate networking between states of examples of positive R&R experience. Actions to mprove Resettlement and Rehabilitation. 4.59 The basic principle of R&R should be to enable those displaced to improve or at least regain the standard of living they had prior to displacement. through restoration of their producive Snua.Six. The objective should be to make displaced persons net beneficiaries of an irrigation development project, and their welfare should be catered for accordingly (Partridge and Salam, 199073). The key actions are74: 72 By contrast, the R&R program under the same project in Madhya Pradesh is currently seriously behind schedlMe. 73 "Land Acquistion, Resettlement and Rehabilitation", W. Partidge and A. Salam, (Irrigation Sector Review Background Paper, 1990). Also, see "Involuntary Resettlement", World Bank Technical Paper No. 80. 74 More detailed recommendations are in Volume II, Chapter VI.  49 (a) Provide land with similar income-generating capacity to that lost. This is the key need, and involves either prior agreement with beneficiary communities on giving up part of their lands in return for the new irrigation service, or provision of government funds for purchase of land from other fanmers. The former may be more appropriate in areas of high population densities such is eastern India. The latter is being used in Gujarat. An additional option for families that express interest in altemative rehabilitation, is to provide them with the choice of reestablishing in an alternative occupation through provision of financial aid, housing and training.75 (b) Adopt a _whole community approach. R&R must have a whole community approach and also cater to the needs of landless, minorities and women. (c) Providre suporting services. A package of supporting services (roads, village water supply, electrification, schools, clinics, training, and other inLrastructure and services) is needed to allow newly resettled communities to reestablish themselves. (d) Establish a Land Acquisition. Resettlement and Rehabilitation Plan, in discussion with the affected communities. Implementation of R&R at the project level is a complex process requiring substantial hands-on involvement of local authorities, community leaders and NGOs. This should have detailed agreed actions, organizational procedures, institutional and staffing requirements and costs. (e) Finance R&R fully as a project component. R&R must be financed not as a supplement but as an integral part of irrigation development and the full funding provided to this project component; and (f) Prepare a National R&R Policy Paper. The above measures can be implemented now on a project-specific basis, but will be facilitated with a clear policy and legal framework. The National R&R Policy Paper, currently under preparation, should become the macro-outline providing general guidance for implementation of R&R at state levels and for GOI approval of specific projects involving resettlement. This should be followed by any necessary enabling legislation at state levels. 4.60 With experience, further improvement of R&R implementation will be possible. To this effect, studies and monitoring of R&R programs are needed. Creation of exactly zomparable sources of livelihood is frequently not feasible, and particular problems emerge in accommodating groups such as tribal peoples where sources of livelihood are diffuse and not perfecdy matched by provision of agricultural land. Nevertheless, while some imperfectons are inevitable, under a we- designed and implemented R&R program material welfare of those displaced can be protected and even enhanced, and this needs to be a standard feature of irrigation development. 75 For inslance, a family may prefer reestablishment in transport or commerce rther than agrculture, but the R&R package for such cases would need careful design to provide sustainable rehabilitation or augmentation of income earning capacity.  50 CHAPrER V. GOVERNMENT/NON-GOVERNMENT ROLES AND CAPAClTY 5.1 Largely as a result of the agricultural sophistication that irrigation has helped induce, management of the irrigation sector has become increasingly complex. Taking yields beyond their present level now requires more exacting water management and a broader range of technical, financial, planning and managerial skills. Not least, the vast scale of operations, maintenance, financing, and general management poses increasing challenges. Yet, irrigation management remains exclusively handled by a governmental structure that has changed little over many decades, and institutional capabilities have declined as a result of limited adaptation and increasing encroachment of political and rent-seeking pressures on sector managers. 5.2 Improving the performance of irrigation will depend most fundamentally on clarifying who does what most effectively, government or the non-government sector; reforming institutional roles as necessary, and enhancing the incentives for and capacity of both government and non- government sectors to perform. For the government sector, there is a major need to reforn management to achieve greater autonomy, accountability and transparency of decision-making and performance. A long-term vision should come into focus, with practical steps implemented for the short- and medium-term in getting there. A. Situation and Trends Dominant Public Sector 5.3 Indian irrigation is dominated by the public sector whose role has increased over time. The scale of most surface schemes in India has necessitated government funding, and similarly, operations and maintenance of these schemes require some public sector involvement. Government management usually extends to the outlet from the minor to the chak (i.e., in most instances, parcels of about 40 ha) and has entailed a large network of grassroots irrigation officials. Collection of water charges also involves a substantial government apparatus. Similarly, with the construction and maintenance of micronetworks, government is becoming a major player. In southern India, management and maintenance of tank irrigation schemes formerly constructed and managed by local communities have increasingly been assumed by the public sector. A notable exception to this trend is the rapid growth of irrigation through shallow tubewells and dugwells where government has played a supporting rather than implementation role, restricting activities to provision of credit, monitoring of groundwater levels and selective investment subsidies to poorer farmers. Nevertheless, even here, where private investment is demonstrably successful, substantial government efforts in eastern India have gone toward development of public tubewells. 5.4 Many functions do need to be handled, or at least steered, by government: policy formulation, planning, design and supervision of public works, legislative and regulatory functions, and some monitoring and support services such as extension and research. Yet, continued reliance on government for sector implementation carries a high fiscal cost, putting the government budget under heavy strain, and diminishes effectiveness in face of the increasingly difficult task of managing infrastructure which is vast in scale and government units that are inadequately structured and trained to meet changing conditions. 5.5 India relies much less on farmer and local community initiatives, autonomous bodies for scheme management and involvement of private entrepreneurs, consultants, and other non- govemmental organizations than many other countries, where they have proven to be a source of vitality and support. Their collective resources should be tapped, and this overview of sectoral performance points to action beyond improvements of existing government institutions to invigorate the sector. Central administrative structures should, as far as possible, be divested to more autonomous local bodies accountable to irrigation's users and with user participation in  51 management.76 All practical possibilities for divesting government investment or implementation to the non-government sector should be pursued. As discussed below, large opportunities are available for divesting at least part of irrigation implementation and management to farmers, entrepreneurs, consultants, NGOs and other government or semi-government bodies. Opportunities are also available for encouragement of investment by farmers and entrepreneurs; for instance, in groundwater investment and development and maintenance of micronetworks. State lrrigation Departments 5.6 Implementing Agency. India's entire surface irrigation and drainage network is primarily managed by state irrigation departments (IDs). IDs handle the planning and design of irrigation and drainage works, supervise contractors during construction, and are responsible for operadon and maintenance of completed schemes, as well as monitoring and financial management. The capability of each ID is the principal determinant of success or shortfall in the state's irrigation sector. Although some states manage to do a largely exemplary job (e.g., Punjab and Haryana), performance of most IDs is weak. The management task is formidable, given present structures and capacity. Many IDs manage an aggregate irrigated area that is larger than that of most countries (Vol. II, Annex Table 7.2). Decentralization and actions to improve management can help, but bureaucratic constraints and inherent disincentives are less easily tackled. 5.7 Intives. IDs and government command-level structures such as CADAs suffer from weak incentives to perform. Polidcal pressures are a large part of the problem. They influence decision-making favoring new constructions. Creating unproductive jobs for constituents is tolerated, even encouraged. Staff rewards and senior staff appointments are often biased in the direction of compliance with such expediencies. Yet more damaging is the growth of rent-seeking influences. Salaries are modest and salary levels and promotions are minimally affected by performance. By contrast, large financial gains are possible through collusion with contractors or illicit collection of revenues from farmers in exchange for water distribution favors. Such practices have become increasingly common and critically alter incentives; they discourage efforts to improve quality control of both construction and maintenance, to make water management more equitable and to improve cost recovery. Good monitoring and sound financial management also carry little reward as they serve to expose these deficiencies. 5.8 Agcountability. These pressures are compounded by a marked lack of accountability. While ID budgets are approved by state finance departments, and central government approves investments involving central Plan funds, in most states these have proven to have little influence on performance levels and even on the use of funds. Such monitoring is restricted to desk auditing of accounts. There is little outside monitoring of field performance by state governments, insufficient requirement for publicly available reporting, and as farmer participation is minimal, little effective means for users to put pressure on IDs to improve performance. A fundamental need is to increase the accountability and transparency of decision-making in the sector, to other government institutions, to the general public and, in particular, to the users of irrigation. 5.9 Structure and Management. Several managerial problems in IDs relate to their organizational structures, staffing and training, which are mainly responsive to construction engineering needs (Frederiksen, 199077). Most IDs are organized on the basis of geography, managed downward through successive supervisors to "circles," each managed by an engineer having responsibility for all irrigation activities within the area. A large number of entities reports 76 "Need for a Strengthened Private Sector in Irrigation: The Service Industry and the Users." H. Fiederiksen (Iigation Sector Review Background Paper, 1990). 77 "Staff Capability - Personnel Policies, Specializadon and Training," H. Frederiksen, (Irrigation Sector Review Background Paper, 1990).  52 to the state Engineer in Chief and Secretary of Irrigation at the top of this pyramidal structure.78 5.10 Meanwhile, irrigation development and management have become more complex. Site conditions are usually more difficult, more demanding design and construction techniques and materials are used, planning must encompass increasing water scarcity, and the demands of multiple use require more thorough and holistic analysis for a whole river basin. Progressively, the operations and maintenance function has become more important as the stock of irrigation infrastructure has grown. In most states, especially those where water development is nearly complete, it is now the dominant functional need for irrigation management. 5.11 Changing nceds are taxing staff capabilities, limited almost exclusively to general engineering. More skills in specialist aspects of construction are required, and staff with expertise in water resource planning, hydraulics, water management, irrigation techniques, and maintenance are particularly scarce. As important, non-engineering skills need incorporating: business planning and management; cost accounting and financial analysis; monitoring and economic analysis; and socioeconomic and environmental considerations. 5.12 Staff Training. Opportunities for career specialization, on-the-job training, and cross- fertilization across states and with other countries are limited. There is no induction training for engineers entering the irrigation service and scant skill development or refresher training is offered during an engineer's career. Exposure to ideas outside the state or to irrigation management techniques and research in other countries is rare. These problems are exacerbated by the practice of transferring staff to cifferent jobs in the ID at frequent intervals, resulting in lack of career development in any one field and also in serious disruptions of work programs. In most states, while exceptional individuals exist, the average professional quality of irrigation engineers has probably declined over time, and there is a marked absence of the non-engineering skills discussed above. 5.13 Since the late-1970s, Water and Land Management Institutes (WALMls) have been set up in 10 states to provide general training to ID engineers in water agronomy and management and some engineering to agriculture department (AD) staff involved in irrigation. The need for WALMIs remains evident, but their effectiveness is hindered by a lack of focus on training needs and a too-theoretical approach to what are essentially practical technical problems (para 5.34). Coordination 5.14 In most states, there is a need to define better the roles of different institutions involved with irrigation and to ensure that functions sometimes falling between them are effectively carried out. Construction and general agricultural extension are clearly defined as functions of the irrigation and agriculture departments, respectively, but other functions are either not assigned or, more commonly, are assigned without staff expertise and resources to handle them, nor incentives to perform the tasks well (see below). Even where roles are clear, coordination among government depattrnents is often weak. 5.15 Extension. Important extension functions are deficient. Water agronomy is neglected, because agricultural extension staff lack training in the optimal use and timing of water. Fanner knowledge is notably weak in this area and is partly responsible for a tendency for head-reach farmers to overwater their crops. Support for farmer participation in irrigation is also weak. Little extension advice is provided to farmers on how to form water user associations and the benefits thereof, as this objective is seldom defined as a priority for one or other of the government agencies. Finally, limited contact among irrigation departments, farmers and agriculture departments has often reduced the effectiveness of surface irrigation design and subsequent 78 ID orgzational restructuring to create functional specializafion is underway in Punjab; for description refer to Second Punjab Irrigation and Drainage Project Appraisal Report (Cr. 2076/Ln. 3144).  53 operations in meeting agricultural needs. Scope usually exists for improved timing of water delivery, based on regular in-season meetings involving farmers and agriculture officials. 5.16 Major shortfalls occur in the construction and maintenan. ; of micronetworks (watercourses, field channels and field drains) and the efficiency and equity of water distribution at that level. Irrigated area, productivity and equity are thus substantially reduced. Government efforts have primarily emphasized government-executed construction of micronetworks, either by the ID, engineering wings of AD, or by the Command Area Development Authorities (CADAs), at the expense of farmer involvement. Resource constraints have limited the number of micronetworks constructed.79 Further, farmers assume that maintenance of government- constructed works should be done by government and maintenance has been poor. Also, little advice has been provided to farmers on water management at the micronetwork level, a common deficiency except on the wel'-established warabandi schemes in the northwest. A larger, more sustainable and much less costly impact would have been achieved if farmers had been assigned these tasks and encouraged to feel they own and are responsible for their micronetworks. The government role should be revised to a strong supporting function: topographic surveys, layout design, financial assistance for construction (credit and perhaps matching grants), assistance in forming water user associations and advice on water management and maintenance subsequent to construction. However, only a few pilot efforts have been made in this direction. 5.17 CADAs. A major government initiative aimed in part at responding to these problems was the establishment of Command Area Development Authorities (CADAs). The CADA program began in the mid-1970s, and there are now 54 CADAs covering 131 command areas representing a total command area of about 18.5 million ha. CADAs were intended to address several of the problems discussed above, notably construction of micronetworks and coordination of different government services involved with irnigation. CADA structures and functions vary widely across states. In some states, the CADA is essentially an executive committee intended to coordinate the activities of different government departments. In practice, many have not succeeded, because of conflicting instructions from line department heads. In others, staff have been seconded from different line departments to CADAs, becoming additional implementing agencies. However, with implementation responsibilities often confined to government construction of micronetworks, such CADAs have in effect become specialized construction departments. For most CADAs, the key problems remain: still-weak coordination between irrigation and agricultural services, limited (and costly) progress with construcdon of micronetworks, poor maintenance and water management on such micronetworks, still-weak extension for irrigated agriculture needs, and limited efforts to generate fanner involvement. A basic problem is that CADAs remain as government departments and have had difficulty in improving on the performance of other government departments that they were intended to supplemnent. This existence has seldom compensated for weak performance of IDs and other departments involved with irrigated agriculture. For most CADAs, unless they can be transformed to become managerially and financially autonomous scheme management entities, there is no strong case for retaining them. Farmer Involvement 5.18 While central government policy statements have supported farmer involvement, little has actually been done at the field level, and much of wham has been done has been promoted by NGOs. Indeed, field experiences have often been the reverse, with incre.sed government assumption of implementation in states such as Haryana (micronetwork eonstruction and maintenance) and in several eastern states regarding public tubewells. This contrasts with the historical development of irrigation which relied substantially on local community investment and operations. The current situation and trends need to be reversed and much can be learned from past ,xperience. There are also a number of recent pilot efforts to learn from (Datye and Patil, 79 Micronetworks for 0.5 million ha were installed in 1989, a fall-off from earlier development in the mid- 1980g.  54 1988), as well as experience from cooperatives in other sectors in India and from water user associations (W1TAs) in other countries. Experience is mixed but provides sufficient guidance for implementation of government actions to promote WUAs (Bottrall and Kathpalia, 1990).80 An encouraging recent development is the much greater interest in promoting WUAs now found at state-levels. For instance, piloting of WUAs has been underway for several years in some states (e.g., Tamil Nadu and Maharashtra), and West Bengal has started transferring government tubewells to Panchayats. GOI intends major emphasis on this under the Eighth Plan and mechanisms to support state efforts for an increased role for farmers are currently under consideration. This will require a progressive and grassroots approach with features as discussed at para 5.46. B. Using the Non-government Sector 5.19 A first and fundamental need in sectoral management is to reconsider and decide on the appropriate role of the public sector in irrigation development. The considerable initiative and dynamism of farmers, entrepreneurs, consultants, universities, research institutes, and NGOs are untapped. Further, government can no longer afford to shoulder the cost, especially at inefficient levels of expenditure, of the sector's development. Government should retain certain responsibilities (para 5.4) but increasingly it should devolve major portions of irrigation investment, operations and maintenance to the non-government sector, including farmer beneficiaries. Private Investment Opportunities 5.20 Groundwater Development. The two largest opportunities for private investment are in groundwater development and development of micronetworks for surface irrigation. Groundwater development is already primarily in private sector hands, with government playing an important supportive role. While inadequate pricing of energy remains a serious issue, in other respects the rapid expansion of private groundwater irrigation in the past three decades is a major success story for the Indian irrigation sector. By contrast, the public tubewells in eastern India (Uttar Pradesh, Bihar and West Bengal) have proven costly, difficult to maintain, to have generally lower yields and have involved particularly high subsidies. Future investment should be focussed exclusively on private development of wells, with government back-up in the form of groundwater monitoring, planning of surface irrigation development to facilitate conjunctive groundwater exploitation, and provision of policy and legislative guidance, research, extension and access to credit. Where deep tubewells are required and would not deleteriously affect groundwater levels and downstream surface flows, private investment in them can be encouraged. Collective ownership by small groups of farmers, using smaller capacity pumps, might be feasible. For overly large groups, social organizational problems could hinder collective ownership. Alternatives are investment by wealthier farmers or entrepreneurs and water selling, a practice that has become increasingly common with the small shallow tubewells. 5.21 Micronetwrzks. Farmer participation in surface irrigation, especially to maintain micronetworks, is a missed opportunity. An alternative to the public sector is required, but first the constraints limiting farmer involvement, in part induced by the politcal environment, must be tackled. Politicians have frequently encouraged an attitude among rural constituents that government assistance is a right. More specifically, little effort has been made by IDs or CADAs to establish WUAs, yet some form of community structure is essential for farmer involvement in micronetwork construction and for subsequent successful tc.eration and maintenance by farmers. In Pakistan, micronetwork development is a farmer respon;ibility with government providing 80 "Farmer Participation in Irrigation Sector Management," A. Boaml and G. Kathpalia, (Draft, 1990), (Irrigation Sector Review Backgmound Note).  55 technical assistance for micronetwork layout and construction. Farmers are required to establish themselves in water user associations before they are eligible to receive credit. Such measures might work in India. Opening of a minor could be conditional on completion of WUA formation and micronetwork construction for a majority of the chaks, and opening of an outlet from a minor conditional on comple-ion of chak infrastructure. Existing government services could be adjusted to provide a strong technical assistance input and ready access to credit and perhaps matching grants.81 Creating Autonomous and Accountable Institutions 5.22 Alternative Structures. For all surface irrigation and drainage commands, the ultimate objective should be the creation of managerially and financially autonomous entities responsible for water management anid maintenance and answerable to irrigation's users. Examples in other sectors in India and of irrigation in other countries pertain. Institutional alternatives to government departments in India include registered societies, boards, public utilities, cooperatives and mixed private/public ventures. These entities have greater freedom and flexibility than government departments, and comparatively greater public accountability due to established open auditing and review procedures. In irrigation, entities such as the Bhakra-Beas Management Board, Sardar Sarovar Narmada Nigam Ltd, several WALMIs and CADAs organized as registered societies, and the Damodar Valley Corporation have demonstrated that creation of alternative structures adds flexibility and dynamism. Wherever possible, such options should be further exploited. 5.23 International ExamDles. In China, for example, irrigation investment and management responsibilities are decentralized. The starting point is river basin administration which transcends other boundaries for aggregate planning and monitoring of water resources. For a major command, a command-level authority confines its role to major structures and canals. Implementation is decentralized through contractual agreements between successive layers -- command, prefecture, counties and village groups. Each level is expected to carry out the works assigned to it, to reimburse loans received, and to ensure cost recovery from the level below. Water charges are used within the irrigation scheme for reimbursement of investment loans and funding of maintenance and adminstration, with agreed percentages taken by each administrative level. At the base, farmer implementation is strong. Accountability at each level and the discipline imposed by clear designation of responsibility and financial autonomy encourage frugality and efficiency. 5.24 In Korea, the center's role in implementation is restricted to planning and monitoring of water resource development by the ministry, and construction of major schemes by the Agricultural Development Corporation. The corporation operates each scheme for two years and then hands over operation and maintenance to farmer-managed boards -- farm level implementation agencies (FLIAs) -- which employ professional managers and fund themselves through water charges. Jaganese irrigation is managed by farmer cooperatives which secure funding and employ contractors for irrigation expansion. Countries such as Mexico and the Philippines are moving towards the features found in China and Korea. Under the Mayo project in Mexico, irrigation districts formerly managed by the state are being handed over to farmer organizations responsible for blocks of about 3000-5000 ha, following successful experience with this approach under earlier projects. In the Philipines, minor schemes are managed by farmers and central investment assistance for expansion or improvement is conditional on the existence of an irrigation association and farmer participation and contribution to investment. For large schemes, starting in the mid- 81 Reservations have been expressed by state and cental governments that this ideal may not be attainable, at least in the shon term. Nevertheless, there is a consensus that farmers should have responsibility for operation and maintenance of micronetworks and greater involvement in construction (active participation in design and at least some labor and financial contribution to construction). This would create a sense of community ownership of the micronetwork, important for subsequent O&M by farmers, and enable progressive reduction in public investment costs.  56 1980s, a phased process for eventual handover to farmer management has commenced, starting with farmer participation in maintenance, revenue collection and operations, and leading to water fees retained at command level and managerial and financial autonomy. The transition is being promoted through strong government technical assistance and training. 5.25 Modelling S The common features of these successes are: a reduced role for government administration; significant participation, management and often owvnership by farmers; and scheme-level autonomy, including the generation of funds. Similar arrangements are recommended for India, but implementation will need practical steps to progress from the existing situations. One opportunity is presented by CADAs. Successful CADAs could be given more independent statutes (as registered societies or boards as found in some states), allowed to retain water charges, and given a phased but finite time period for progressively reduced government financial contributions and increased farmer participation in management. The hand-over would need a strong extension effort to create WUAs, and successive tiers of farmer representation in higher-level management, accompanied by a phased program to reduce government staffing and to place remaining staff under CADA authority employed and answerable to the farmers. Scheme infrastructure should also be repaired to good condition as part of the hand-over agreement. In states such as Kerala, CADAs are already more autonomous than elsewhere and this transition could be more rapidly implemented. 5.26 Another opportunity is presented by existing institutions or where new institutions are desirable. The Bhakra-Beas Management Board (BBMB) usefully coordinates water distribution among Punjab, Haryana and Rajasthan. In many river basins, there is also a need to create institutions for managing water planning and distribution between states and users. These should be created on a commercial basis, as far as possible, with clearly defined fees, rights and obligations for each user entity and the overall basin authority. During the planning of new schemes, the eventual institutional structure and financial arrangements should be agreed with users from the outset. Such innovations will take time to become widespread, but opportunities exist in all states to implement pilot developments. 5.27 rmtiactors. Consultants. Universities and NGOs. In parallel, large opportunities are available now to IDs to promote increased private sector participation. Contractors can play a much larger role in irrigation development. For maintenance, use of force account is still common. Yet under existing law each laborer or foreman hired by government tends to become a pernanent annual wage and pension obligation, adding to recurrent costs. Maintenance should be handed over to contractors, with the scheme manager's role as a monitor and enforcer of maintenance standards, including independent outside monitoring where scheme management is still under ID adminstration. 5.28 To date, the use of private consultants has been limited. The irrigation-sector consulting industry is in the early stage of development, but consulting capabilities have grown over time through the efforts of retired engineers, and a large consulting capability in other disciplines already exists. GOI and IDs should more actively enhance their technical expertise by using consultants in needed areas without adding to the permanent payroll. Increased use of consultants may add to costs in the short-term but provides critical skills for a defined period, enhancing in- house capabilities to implement a task requiring specialist input without permanent wage-bill implications. A large part of the consulting input is needed in fields such as management information and monitoring systems, cost accounting, economic analysis and non-engineering aspects of project preparation. Foreign consultants could also be used more actively. This recourse is rarely employed in India due to foreign exchange considerations. Yet few, if any, countries are in a position to develop a complex sector without use of international consultants. For major projects it is generally essential and is proving an important input in the Sardar Sarovar Project.  57 5.29 The resources and initiative of universities, research centers and NGOs can also be contracted more widely. A network of universities and research centers exists in all states and has in a number of cases been used effectively to support irrigation development at relatively low cost. These institutions could be used more extensively to undertake monitoring and evaluation, baseline surveys, some technical research, part of the in-service training of ID) staff and assist with irrigation project preparation. Scope also exists for more active use of NGOs which have been effectively used in irrigation activities requiring a grass roots involvement such as promoting WUAs, implementing R&R programs, and catering to the needs of women and minorities. 5.30 Private-sector Labor. Also, the present large network of ID minor officials used for basic irrigation management (e.g., canal inspectors, public tubewell operators, collectors of water charges, Jilladars, Patwaris, etc.) could be eliminated. Most such functions can be better handled by village laborers employed, paid and answerable to WUAs. WUAs would decide on the level of staffing required and would have the ability to discharge personnel performing unsatisfactorily, a feature deprived to government. State back-up could be re,stricted to a good training and technical assistance function. If government control of impletnentation is still considered necessary, contracts can be established with private firms to provide staffing, enabling future flexibility to change unsatisfactory staff or to dispense with suc'i functions. C. Improving Govement Institutional Performance 5.31 While the above recommendations would enable a much larger role in irrigation for the non- government sector, an important government role will remain. Even in the long term, government will need to plan water resource development and coordinate its exploitation (but increasingly through river basin authorities), to monitor sector performance and to channel government funds, including supervision of contractors, for large public works. As transition will take time, I)s will remain primarily responsible for irrigation development and management in the medium term. The following measures are recommended to improve the management capabilities of IDs and to steer them towards their longer-term role. R estructuring Irrigation Departments 5.32 Functonal Specialization. The present structure of IDs, which is inadequate for functions other than construction, should be restructured to create specialized headquarters divisions, typically: operations, maintenance, planning, design, construction, training and staff development, monitoring and evaluation, and administration (Volume II, Figure 7.3). Such functions should also be reflected at the circle level. Many circles would have operations and maintenance as their only function and need to be properly trained and motivated to do this well. Parallel efforts are needed to create greater specialization within the engineering cadre. Particularly for non- construction activities, staff specialization needs to be fostered through carcer development opportunities, reduced transfer rates for staff and, as possible, through special financial incentives for less popular functions. As important, non-engineering skills need to be incorporated (e.g., cost accounting and financial management, monitoring and economic analysis, and socioeconomic and environmental disciplines). Training of existing staff also needs to be intensified and would be facilitated by the training and staff development division proposed below. 5.33 Staff T_aning. Training and access to ideas in other states and countries will form the basis for the future vitality and competence of the sector's institutions. Professional capabilities of staff should be upgraded. Training should also involve "retraining" for personnel in surplus functions to provide them with skills for changing over to functions where ID manpower may be in short supply. WI-ere an ID needs to reduce staff (para 3.14), budgeted retraining using existing state institutions could include retraining for private sector job opportunities in non-irrigation engineering functions. A tnaining and career development division should be established in each  58 ID to provide the necessary management focus on this objective. Using existing training facilities (field training centers and WALMIs), supplemented by contracts with universities and consultants, training targeted toward the priorities established by each ID is also needed. Other state resources include universities, engineering institutes, and various training institutions in management, finance, and economics. A well-used budget for titijizing such centers could provide greatly expanded training at reasonable cost. These initiatives should be supplemented by exchange programs and visits to irrigation projects and institutes in other countries. While the budget for such training and exchanges would be much increased from present minimal levels, it would still be small relative to other budgetary items for irrigation and should not be skimped. 5.34 Some rethinking is recommnended for the WALMI training programs. Training has tended to be too theoretical and has not had much practical impact. Training for ID staff should be adjusted to feature water management, maintenance and design, and include management skills, cost accounting, diagnostic analysis, monitoring and planning. Training in agronomy for irrigation engineers can be kept to the minimum needed for engineers to be able to relate to agriculture staff and agricultural concerns. For the staff identified to provide the technical assistance to farmers for WUA fornation and operations and maintenance of micronetworks (ID, AD or CADA), WALMIs need to provide pracical training for these needs (in community organization, operations and maintenance, basic hydraulics, topographic surveying, water agronomy, and simple construction skills). MOWR is also considering establishing a National Irrigation Management Institute which would promote new initiatives in irrigation and assist the WALM'IS in developing their training programs. 5.35 Reduce Staff Transfers. A well-acknowledged problem is the practice at both ID and central government levels (and at all levels in each hierarchy) of frequent transfer of staff. This creates a serious lack of continuity and accountability. Competent staff should be kept on one assignment for a minimum of 3-4 years (or longer if a task, project or management drive needs seeing through). Promotions and other incentives should be adjusted to reward continuity and good performance. This will nr.d changes from the existing incentives and promotional policies. Special additional compensation (e.g., supplementary housing or transport allowances, other special financial allowances) will be needed for tougher and more remote work situations or less popular functions (e.g., planning and design). Persons continuing in functionally specialized positions will need to have better access to promotions than individuals seeking to rapidly switch posts. 5.36 Improve Staff Accountability and Incentives. Seeing through and successfully achieving a demanding task should be the principal criterion for a promotion. Staff should be made accountable for their task with sanctions applied against poor performance and rewards for good performance. Existing examples of positive incentives in Indian irrigation should be more widely and systematically applied: for instance, use of one-time monetary awards for a well performed task; selection of good performers for further special training, overseas visits and fellowships; special mrnention in the ID's Annual Reports or in AARs (see below), etc. 5.37 Reestablish Annual Administration Reviews. At command or sub-project level, the former practice of producing "Annual Administration Reviews" (AARs) has fallen into disuse or has become a nominally applied reporting procedure without useful application as a management tool for monitoring performance and targeting improvements and next year's budget. Where still ongoing, for instance as done by the Bhakra-Beas Management Board,82 it has proven useful for 82 Past BBMB Annual Administration Reviews not only describe with good engineering details what maintenance, water distribudon achievemnents and other works have been done, but their timing, commendations on exceptionally well performed sub-asks (particularly important for staff morale) and analytical discussion of weak performance and how to improve matters. More cost accounting analysis would have been desirable. Future AARs could include a short executive summary highlighting both matters of pride (good or improving performance), commentary on problems and how to solve them and, in particular, targeted performance improvements for next  59 both command managers and state and central authorities. Government has decided to require reestablishment of this procedure including a stronger emphasis on cost-accounting and staff- requirement analysis. 5.38 Increase Public Involvement and Accountability. In the drive to improve sector performance, public involvement and accountability to the general public should be actively sought for by IDs, senior state authorities and GOI. Approvals of state irrigation programs and monitoring of annual performance should involve closer and more formal interlinkage with other government departments (finance, planning, agriculture, development commissioner, etc.). The general public and the media should also have open access to reports. The Indian academic community should be encouraged to comment on such material and their help actively enlisted wherever useful (e.g., thinking through development and performance improvement priorities, project impact evaluations, R&R programs). The transparency and accountability introduced by this process should help IDs, senior state authorities and GOI to better evaluate sector performance and focus on irrigation development and improvement priorities. 5.39 No Growth in Staff. Reforming IDs and their performance does not imply growth in overall staff numbers. To the contrary, existing staff need to be reallocated, deserving cadres sent to short to medium-term training courses, and consultants used for specialized activities and training. If some additional hiring is needed in non-engineering fields, it can be more than compensated by attrition due to retirement resulting in the net reduction of ID staffing recommended in para 3.14. 5.40 Management Culture. At the heart of such improvement measures is the need to create in each ID a "management culture" incorporating skills and perspectives of management specialists, economists, financial analysts, agriculturalists and other specialties. Above all, skill, motivations, and techniques used in successful private corporations should be emulated. No government entity has the full managerial freedom of a private enterprise, but within existing constraints, much can still be done: establishing priorities and use of staff assignments to achieve these priorities; use of management information systems for rapid diagnosis of problems; establishing accountability for staff and providing rewards related to performance; and use of cost accounting to analyze performance of all key ID activities. 5.41 Political Will. Success will depend primarily on orientation and support from each Secretary of Irrigation and importantly, the support he receives from the state leadership. Wherever a Secretary's decisions are reversed or undermined by political ;nterference, there is little possibility to change matters. This has increasingly been happening, I. to a major degree in some states. Staff appointments, investment decisions, selection of contractors, monitoring of construction quality and other implementation matters, and policies related to cost recovery are less and less governed by objective criteria. By contrast, where political involvement has been judiciously applied in a positive manner, this has proven an invaluable supporting force for change. Coordinating Government Institutions 5.42 The respective roles of IDs, ADs, CADAs, other government institutions and farmers need clarifying in a ntmber of states. IDs need to concentrate on ensuring that the main irrigation and drainage systems operate and are maintained well. Below the government outlet, operation and maintenance of micronetworks should be made a farmer responsibility and their contribution to construction increased. This will only be successful if credit and extension services are provided year. For completed commands, the AARs should also analyze impact (water distibution achievement and fanner yields). AARs should be widely disseminated: to the comn.nd's own staff, to various state authorities, to GOI and to irrigation beneficiaris (WUAs), and should involve a sign-off and supplementary comment prress by ID, finance depatment and water user representives.  60 to farmers to help them establish WUAs and for provision of technical assistance. Either the roles of CADAs need to be adjusted to this function or special extension provided by ADs or IDs. Agricultural extension by ADs also reeds to be improved, particularly in water agronomy. Despite the existence of CADAs, coordination between government departments often remains weak; as done in some states, coordination committees at state, district and scheme level (Volume II, Figure 7.4) could help improve coordination. It is also noteworthy that the states without CADAs -- Punjab, Haryana and Tamil Nadu -- have managed in these and other respects as well or better than other states. A feature of these states is that the IDs have better management than typically found elsewhere. The workable alternative to CADAs has been to concentrate on ensuring that each department (irrigation, agricultare, etc.) fulfills its assigned role and that necessary coordination is achieved. Increased Role for Farmers and Women 5.43 Regarding public delivery of irrigation services, the goal is to assign roles clearly so that necessary irrigation and support functions are performed well, without adding to overall cost and staffing. This necessitates a greater role for farmers. The central and state govemments endorse the view that farmer participation through water user associations (WUAs) or other forms needs to be promoted so that farmers can participate in or assume decision-making, management and implementation of irrigation operations, especially of micronetworks. WUAs provide a local forum for decision-making and resolution of disputes between farmers by the user community as a group. They also provide a community structure for undertaking self-help, such as maintenance, watercourse lining and better water management, based on local knowledge of how the system is working aiid farmers' needs. Farmers should also progressively participate in management at higher levels in the irrigation command; representatives of committees at chak level can form committees at minor level, and subsequently at the whole-system level eventually leading to the longer-term goal of farmer management of autonomous schemes. 5.44 Participation and Benefits for Women and Minorities. Nearly one-half of agricultural labor is female, and over 30% of poor households in rural India are headed by woinen. Similarly, a disproportionate number of poor households are landless families or minority groups such as scheduled castes and tribes. In the case of irrigation, men traditionally make decisions on water allocation and participate in monitoring of usage, and larger families from more influential societal groups have the largest say. Raising the productivity of women, the poor and minority households requires specific features in irrigation project design and implementation to cater for their greater participation (Mitra, 199083): representadon in water-user associations; extension services designed to reach these groups; and access to drinking water, irrigation for communal fodder and woodlots and water facilities for cattle, bathing and laundry. 5.45 Initial project design should include a baseline socioeconomic survey to ascertain the needs of women and minorities. Such actions are increasingly present in projects, often with assistance from NGOs. These initiatives now need to be more systematically applied across the sector as a whole. 5.46 Developing Water User Associations. Promotion, development, and support of water user associations is likely to be the most effective vehicle, short-term and long-term, for effecting farmer participation and control of irrigation schemes. Beginning at the chak level, WUAs can eventually permeate all levels of sectoal management. It is also the first step in localized decision- making and cost recovery that can spiral upward. Simultaneously, scheme-level adminstration can actively reach down to WUAs to enlist them in widening circles of consultancy, participation, decision-making and cost recovery. Successful development of WUAs is likely to depend on: 83 "hnproving the Productivity and Equity of Irrigation Projects trough Integation of Women's Perspectives." M:. Mita, 1990, (Irigation Sector Review Background Paper).  61 (a) Reliable and predictable water slpgly to the head of each minor; under an agreed operation plan; (b) Common economic incentive for farmers to group together to secure tangible and reasonably early gains (c) Defined rights and obligations for government and WUAs encompassing the agreed water delivery service to the minor head, WUA payments for the service, and respective WUA-govemment maintenance responsibilities, including reduced payments where farners handle maintenance and operations; (d) Equitable participation and operation.1 integrity (periodic elections, defined rights, including safeguards to protect the interests of small farmers, minorities and women, a written constitution and by-laws, and regular meetings of the executive and general body84); (e) Assistance of NGOs, private consultants, social activists, and local leadership i, organizing farmers. Diverse disciplines are required: irrigation engineers, sociologists, management and institutional sr -ialists and trainers; and (f) A grass-roots aproach, progressively building farner cap .bilities and responsibilities. 84WUA saute features generally regarded as impoat to successful and intemally equitable WUAs include: regular meedngs of the executive and geneal body, periodic elections of committee officials including rles to limit the prmiissible tenure of committee officials, defined rights and obligations far all WUA members, annual elecdtions for the committee with a vote for each WUA mamber, meaes to ensure representation in decision- making by evay chat and including conmittee participation by minirities and women, and procedures for outside auditing of accounts, all safeguarded under a written WUA constitution and by-laws.  INDIA - IRRIGATION SECTOR REVIEW HGHLIGHTS OF RECOMMENDED ACTION __UATI RE_ fIB3nm TRW FRAME I. FORE CONERENT WATER RESOURCES POLICY Reaffirm the National Water Policy and propare compatible state G01 coordination/state 12 months water policies (COI Plan funding conditionality) governments Establish river besla planning commissions (and imploemntation GOI/state governments Initiate tmmediately euthoritier for multipurpoe /state projects) Strengthen data collection and resource analysis IDa/MO1R agencies Quickly revivw noeds, (increasingly by rivor refl ct in next budget basin planning commissions) Isprove central coordination OlR/other ministries Medium term 2. PRIORITIZE EXPENDIlIRE AND TIoHEN FINANCIAL MANAGEMENT Refocus public lnvestment to prioritize: ID/state governmnt budget Iwmediste reviow and budget -maintenance procoss soplication -Water mnageant -private groundwater dewvlopment (government support to farmers) _-ecronetworks; farmer OM, farmer partieipation in Investment (governmnt support to formrs) -prol et completions (viable) -drainage -(new surface irrigation investmnt. restricted to whore regional develop_nt neds compel) Upgrad project preparation and evaluation capabilities. Prepare a DDn and CWC Initiate imomdiatly piplitne of well-prepred, viable projects. Curb recurrent expenditure growth and financial wasta2e: ID implementation. State Immediste application base -halt staff growth (and use retirement. to redce staffing) finance departments to on quick diagnosis and -establIoh manag em Informtion, expelture analysis and coot review interim-action plan. accounting systemn (enlist consultant expertise) Improve cost recovery: -establish an annual review and assessment mechanism for water Committees appointed by I-mmdite appointment and charge stat. govornments implemetation -reduce government involvement In collection Panchoynts/1JAs Medium term -remove eIectric power subsidy Stat. governmnt d cision Short-term need n > -prpare program to lmplement moetr-based electricity pricing Electricity boards Modiu term = Strengthen central monitoring of state performance and apply central Planning and Finance Implement for next budget _ix funding conditions regarding: Com issions (CWC technical and thereafter -financial disceplino (review state financial reports). inputs) 0 -full funding of mintenance need (state maintonance plan, budget and pertormnc reiwt) -ivestment focus (on prloritlos) -annual cost recovery reporting  ' ' NATION RESPONSIULl TIE FRA_ S. IWUOlE TEtHICAL PERFCRAMiCE Ensure maintinance ax first priority: -plan budget A monitor using an annual ID maintenance plan, IDc (states, 001 to roview) Urgent priority t and pertormnce review -strengten I4 maintenance division IDa -strengthen extension for micronetwork mintenance (by farmrs) IDu, ADs, or CADAs Impl_met nationwide water mnagemnt Improvement program: -prioritise c_mand-by-command diagnostics, operation plans and IDn A CDI sponsorsh'p A Immediate preparation and management and Infrastructurwl improwents funding priority funding -step-up appliod rosearch and networking on Irrigation technology IDc, research centers -soelctlvely pilot more Intensive modernizations IDn -ephanslz agricultural extension on optimal water usage ADs, WALMIe Upgrade quality control of construction: -empasie rigorous on-site management and inspection procedures IDn Sustained drive required, -improve construction dsign initiate lmodiately -tighten contracting procedures -increse accountability (independent auditing) Improve environmental lopact: -prpare national environ_mt policy paper MOR/Mlin.of Environment 12 months -accelerate/integrate environmental assessment in projects State a COI All future projects -upgrade environmentl monitoring and Institutional capabilities States A 001 Progressive Ensure rehabilltatlon of displaced communities: -preparo ntional RNR policy paper 00G 12 months -implement good RNR progras In all project. IDu All future proJocts 4. EMNWCE DISTITUTIONAL CAPABIUITIES AND ROLE OF PRIVATE SECTOR Exploit private sector capabilities: -increase us of contractors, consultant., universities, research IDa 12 months conters and NQDs. -support groundwater end micronetwork investment by farmer*, IDa (or CADA, AD for 12 months (initiate) enhance g verne_t credit, extension, and support to croate IUAs nicronotworks) Create autonoous surface Irrigation commands: -identify sultable cmmands for piloting IDs 12 months (identify), then -devise & implement progressive handover to users (statutes, Suitable CADAs or other implement :J retention of water charges, UA*, and manag emnt by farmers) 0 Restructure and strengthen IDe IDs Initiate lindiately, major x -create functionally specialized divisions sustained emphosis 0 -Inject mon-engineering skills -upgradestaff training -broaden manaageont focus  Annex 11 INDIA IRRIGATION SECTOR REVIEW List of Review Background Papers Barber, W., Patel, C.C., et al, "Assessment of Indian Water Resources and Irrigation Development Potential," (1990). Baudelaire, J-P., "Design and Management of Future Irrigation Systems to Optimize Agricultual Productivity," (draft 1991). Berkoff, J., "Irrigation Management on the Indo-Gangetic Plain," (1990). Berkoff, J., "Past Lessons for Future Bank Strategy in the Irrigation Sector," (draft 1990). Bhatia, R., "Water Pricing and Cost Recovery in the lIrigation Sector," (draft 1990). Bhatia, R., "Irrigation in India's Agricultural Economy," (draft 1990). Blinkhom, T., Rees, C., "Environmental Issues Related to the Irrigation Sector," (1990). Botrall, A. and Kathpalia, G., "Farmer Participation in Irrigation Sector Management," (draft 1990). Frederiksen, H., "Need for a Strengthened Private Sector in Irrigation: The Service Industry and the Users," (O990). Frederiksen, H., "Staff Capability - Personnel Policies, Specialization and Training," (1990). Mvitra, M., "Inproving the Prductivity and Equity of Irrigation Projects through Integration of Women's Perspectives," (1990). Partridge, W., Salam, A., "Land Acquisition, Resettlement and Rehabilitation," (1990). Pike, J., "Implementation Issues for Water Management," (draft 1990). Price, W., Malhotra, R.K., and Fauss, G., "Improving Construction Quality of India Irrigation Civil Works," (1990). Price, W., "River Basin Planning - Background Note," (1990). Singh, S., "Notes on Water Management," (1990). Smedema, L.K., "Land Reclamation and Drainage," (1990).  Annex m Page 1 of 7 INDIA . IRRIGATION SECTOR REVIEW Bibliography Albinson, B. and Berkoff, DJ.W. (1988). Water Resources Management: Meetingr the Water Demands for Irigantion and Other Comnetina Secto. (From the World Bank Ninth Agricultural Symposium: Innovations in Resource Management). Anker, R., Khan, M.E.and Gupta, R.B. (1988). Women's Particioation in the Labor Force, ILO, Geneva. Asian Development Bank, International Irrigation Management Institute, (1986). Irrigation Service Fees. Proceedings of the regional seminar on Irrigation Service Fees, Manila, 21-25 July 1986. Barber, W.M. (1985). The Rising Water Table and Development of Waterlogging in Northwestemn India. South Asia Projects Department, Irrigation II Division, World Bank. Berkoff, DJ.W. (1988). Irrigation Management in South India: The Anoroach of the National Water Maneet ect. ODI/IIMI Irrigation Management Network Paper 88/2d. Berkoff and Slade, R. (1987 draft). perational Guide to Monitoring paigation Water Managment in India. The World Bank. India Departnent. Agriculture Division. Bhalla, G.S., G.K., Chadha, S.P. Kashyap and R.K. Sharma (1990). Auctal Growth and Structural Chanres in the PRniab Ecnomy: An Input-Output Analysis. EPR Research Report No. 82. Washington, DC: IFPRI. Bhalla, G.S. and D.S. Tyagi (1989). Pattems in Indian Agricultural Development - A District Level Study. New Delhi: Institute for Studies in IndusWial Development. Bhatia, Ramesh (1989). -inancing Irrigatiol Services in India: A Case SVU of BH= states. Binswanger, Hans P., Shahidur R. Khandker and Mark Rosenzweig (I 89). The Imnact of Infrastructure and Financial Institutions on Ag, ultural Output and Investment in India WPS 163. Washington, DC: World Bank. Bos, M.G. and Wolters, W., (1989). Project or Overall krigation Efficiency. Proc. Inl. Symposium on Irrigation Management, Southampton University, UK. Brown, Lester R., (1990). State of the World: A Worldwatch Institute Report on Progess Toward a Sustainale SocitY. W.W. Norton & Company, Inc., New York City. Burt, Charles M. and Plusquellec, H. (1989). atIr DeliveCry.on1. Chapter 11 of On-Farm Inigation Management by the American Society of Agricultual Engineers (1990).  Annex ml Page 2 of 7 Central Water Commission (1988). Water Resources of India. CWC Publication No. 30/88, New Delhi. Chambers, R., Pacey, A. and Thrupp, L. (1989). Famer FirsL Farner Innovation and Agricultral Research. Intermediate Technology Publications, London. Chambers, R. (1983). Rural Development: Putting the Last FirsL Chambers, R. (1988). Managing Canal Irrigation. Practical Analysis from South Asia Institute of Development Studies, Univ. of Sussex U.K. Ox.ford & IBH Publishing Co. PVT. Ltd. Chandiwana, S.K., et al. (1988). Control pf Schistosomiasis Transmission in Newly Established Smaliholder Irrigation Schemes. Chitale, M.A. (1991). Water Resources Mangement in India: Achievements and Perspectives. International Workshop on Water Resources Management, June 1991. Chopra, K., (1989). Land Degradation: Dimensions and Causalitie. from India Joumal of Agricultural Economics, Vol. 44, No. 1, Jan.-Mar. 1989. laines, S. & Pawar, J. (April 1987). Economic Returns to Irrigation in India. Report to Agency for International Development. Dandekar, V.M. (1967). Presidentialddress: Planning in Indian Agriculture. Indian Journal of Agricultural Economics, Vol. XXII, No. 1, Jan.-March 1967. Datye, K.R. (1990) Participative Water Manaement in Irrigated Areas. A Prsal for Action. Centre for Applied Systems Analysis in Development, Bombay. Datye, K.R. (1989). Farmer Participation in Water Management in the Command Area. Centre for Applied Systems Analysis in Development, Bombay. Davidson, G. (1983). de Janvry, A. and Subbarao, K. (1986). Agdcultural Price Policy and Income Distribution in India. Oxford University Press, Institute of Economic Growth, Delhi. Des Bouvrie, C. (1989). Irrigaton Mnagement in India; Thn. Now and in the Future. WAMANA, October 1989. Desai, Thiru C. G. (1989). Critical Issues of Imrigad Agiicule in India, From the 13th Dr. Visvesvaraya Memorial Lecture. Madras, September 28, 1989.  Annex HI Page 3 of 7 De Wolfe, J. (ed) (1987). Jiigatlon and Hath. International Land Reclamation Institute, The Netherlands. Dhawan, B.D. (1983). Development of Tuuewell Irrigation in India. Agricole Publishing Academy, Insdtute of Economic Growth, Delhi. Dhawan, B.D. and Sai, K.J.S. Satya (1988). Economic Linkages Among Inigation Sources: A Study of Beneficial Role of Canal Seepage. Institute of Economic Growth, Delhi. Dhawan, B.D. (1988). Irrigation in India's Agricultural Development: Productivily. Stability. Equity. Institute of Economic Growth, Delhi. Dhawan, B.D. (1989). Studies in Irrigation and Water ManagemenL Institute of Economic Growth, Delhi. Frederilsen, H. (1989) Are the Non-Technical Deficiencies Overwhelming the Technical Advances in Irrigation? Frederiksen, H. (1987). Irrigation and Drainage Systems - Public Utilities That Are Too Often Imrnerly Oriented and in Need f Strenghened Capability. Xlll ICID Sept. 14-26, 1987, conference. Frederiksen, H. (1988). Irrigation: Issues. Policies and Lending Sagies. Paper presented at the 8th Agricultural Symposium, T-he World Bank. January 6-8, 1988. Frederiksen, H. Water Resources Institutigns - Backgond Discussion. Presented at the Seminar "The Role of Institutions in Water Reso'.ces Development and Management," April 4-5, 1989, World Bank, Washington, D.C. Goldsmith, H. and Makin, I.W (1988). A Stud of Water Distrion Performance Under Waradi in Punjab. dia. Draft Report OD 103. Hydraulics Research, Wallingford. Govemment of India (1989). AgrnClimatic Regonal Planning: An Overview. Planning Commission, New Delhi. Government of India (1985). Re=ot of the Study Group on AIricultma1 Strategies for Eastern Region of India. Planning Commission, New Delhi. Government of India. A a u f f , (various years), Ministry of Agriculture. Government of India. Dtorte of Economic and Statistics Various statistical reports, Ministry of Agriculture. Government of India, (1980). National Peetves forWaterResources De New Delhi.  Page 4 of 7 Govemment of India, (1987). Natnal Water Policy, Ministry of Water Resources. Government of India (1988). Rates for Surface Water in India. January 1988, Centml Water Commission. Government of India & USAID, (1987). Water Resources Development. Environmenl Issues and Solutions. Proceedings of Northem Regional Seminar, 24-25 September, 1987. Grimshaw, R. and Greenfield, J. (1987). Vetiver Grass. A Method of Vegetative Soil and Moisture Conservation. P.S. Press Services, N. Delhi. Hazell, Peter B.R. (1982). Instability in Indian Foodgrain Production. IFPRI Research Report No. 30. Hazell, Peter B. and Haggblade, Steven (1990). Rural-Urban growth Linkages in India Background paper prepared for the 1991 Country Economic Memorandum for India, World Bank. International Irrigation Management Institute (1987). Public Intervention in Farmer-Managed rrigation Systems. International Rice Research Institute, 1986. WYorld Rice Statistics 1985, Manila. Kahnert, Friedrich (1989). Assisting Poor Rural Areas Through Groundwater Irrigatiop, Exploratory Poposals for East India. Bangladesh and Nal. World Bank, Asia Regional Series, Report No. IDP 44. Kathpalia, G.N. (1988). Command Area Develoment Programme in India: Prress and Future. Keynote Paper, Seminar on Command Area Development Program, October 28,1988 at Anaimalai. Malhotra, J.R. - chairman (1989). eport of Workinz Group on Irrigation Management ImDrovement Programme for the Formulation of Eighth Five-Year Plan (1990-5) set up by Planning Commission. GOI. Malhotra, S.P., Raheja, S.K. and Seckler, D. (1984). A Methodology for Monidorig Performance of Larg-scal Irrigation Systems: a cas study of uiie YVaabandi Syst of Northwest Indi Agricultural Administradon. Vol. 178, New Delhi. Maudgal, Dr. S. Relevance of Environmental Aspects for River Vaft Deelopent Poects. from Guidelines for Environmental Impact Assessment of River Valley Projects, GOI. Mehra, S. (1981). Instability in Indian Agriculture in the Context of the New Technology, IFPRI Research Report No. 25. Ministry of Agriculture (1987). Reo of the Task Force 1987-88 on Optimal Utilisation of Water Resources in Reservoirs. Department of Agriculture and Cooperation, New Delhi.  Annex II Page 5 of 7 Murthy, Y.K. (1988). Design Enginecring and Consultancy Services for Water Resources Development in India. World Bank & National Hydroelectric Power Corporation, New Delhi. Oblitas, K. (1989). Prospects and Issues for Private Sector Development in Indian Agricul=t. World Bank. O'Mara, Gerald T. (1988). Issues and Options in IrriEition Finance. World Bank. O'Mara, Gerald T. (1988). F.Jciency in Irrigation: The Conjunctive Use gf Surface and Groundwater Resources. World Bank. Pachori, R.S. Farmer's Involvement in Irrigation Management Patel, C.C. (1989). Water Availability and UIse. from the Third Indian Engineering Congress, Madras, The 14th Bhaikaka Memorial Lecture. Patil, R.K. (draft 1989). Farmer Participation in Irrigation Management. Maharashtra State. Centre for Applied Systems Analysis in Development, Pune. Perry, C.J. (1989). Irrigation Sector Position Paper (draft), India Agriculture Operations Division, World Bank. Perry, C.J. (1991). Irrigation Scheduling and the Productivity of Water in Water-Short Indian Irrigation Systems. (Draft Mimeo) Plusquellec, H. (1988). Improving the Operation of Canal Irrigation Systems. Economic Development Institute and Rural Development Department, The World Bank. Plusquellec, H. (1989). Two Irrigation Systems in Colombia. Their Performance and Transfer of Management to Users' Associations. Agriculture and Rural Development Dept.-The World Bank. WPS 264. Plusquellec, H. (1983). Land Consolidation in Irrigation Projects. Journal of Irrigation and Drainage Engineering, Vol. 109, No. 1, March, 1983. Price, W. and Barber, W (draft 1990). Water Resources of the Ganges Basin. An Assessment of their Development. Availability and Demands. Agriculture Operations Division, India DepLfTechnical Dept., Asia Region, World Bank Report 9127-IN. Radhakrishna, R. and Ravi, C. (May 1990). Food Demand Projections for India, Center for Economic and Social Studies, Hyderabad. Study commissioned by India Agricultural Operations Division, World Bank.  Annexl]1 Page 6 of 7 Ramamurthy, " (1987). impact of Irrigation Projects on Women. USAID. Ramanujam, K. Water ManaRenient in the Various States - An Overview of the Problems & Action Requirements, Ministry of Irrigation. Rath, Nilakantha (1989). AgricultualGrowth and Investment in India. Journal of Indian School of Political Econ2ny. (January - June) pp 64-74. Repetto, Robert (1986). Skiming the Wpater - Rent-Seeking and the Performance o IPublic Irrigatiin Schemes. World Resources Institute. Reserve Bank of India, (1984). Agncultural Productivity in Eastern India, Report of the Committee on Eastern India. Satpathy, T. (1984). Irrigation and Economic Development. Ashish Publishing House, New Delhi. Shah, R.B. (1989). Role of Maior Dams in Indian Economy. Central Water Commission. Singh, Satnarayan (1989). Warabandi - An Overview. WAMANA, October 1989, Pages 6-9. Singh, Satnarayan (1989) Issues in Canal Operations in South India: PasL Present and Future WAMANA, July 1989, Pages 1-13. Sinha, B and Bhatia, R. (1984). Economic Appraisal of Irrigation Pro&ects in India Agricole Publishing Academy, Institute ef Economic Growth, Delhi. Small, Leslie E. et al., (1989). Financing Irrigation Services: A Literature Review and Selected Case Studies from Asia. International Irrigation Management Institute, Colombo, Sri Lanka. Svendsen, Mark, (1988). Sources of Future Growth in Indian Irrigated Agriculture. Paper presented to the planning workshop on Policy Related Issues in Indian Irrigation, Ootacamund, Tamil Nadu, April 26-28, 1988. Org. by Internatl Food Policy Res. Inv. Vaidyanathan, A. (1987) - Madras Institute of Development Studies. Irrigation and Agricultural Growth. Presidential Address to 47th Annual Conference of Indian Society of Agricultural Economics. Vaidyanathan, A. (1988). Critical Issues Facing Indian Irrigation. Paper for workshop on "Pclicy Related Issues in Indian Irrigation," April 1988. Wade, Robert, (1988). Village Republics: Economic Conditions for Collective Action in Scuth India. Cambridge: Cambridge University Press, 1988.  Anna TR Page 7 of 7 World Bank 1991 Country Economic Memorandum (Vol. II): Agriculture: Challenges and Opportunities, Report No. 9412-lN, Agriculture Operations Division, India Department. (1991). Agicultu-al Resath in-India: Prologzue Performance and Prospects. Report No. 8383-IN. India Agriculture, Operations Division. (1990). Annual Review of Evaluation Results for 1989. Agriculture and Rural Development Sector, Operations Evaluation Department. (1990). Asia Water Resources Study: (draft annexes on Indonesia. Philippines & India, Asia Region Agricultural Technical Division. (1991). Database on Irrigation Projects Audited. 1974-1988. Operations Evaluation Department. (1990). India: Agricultural Extension Review. Report No. 6805-IN, india Agriculture Operations Division. (1987). India: Agricultural Technology Review. India Agriculture Operations Division. (1989). India: Povert. Employment and Social Services. Report No. 7617-IN, Country Economic Memorandum, India Department. (1989). India's Environment: A Strategy for World Bank Assistance, Report No. 7676-IN, India Department. (1989). India: Strategy Paper for Rural Water Supply and Sanitation. (1989). Planning the ManaeemenL Operation. and Maintenance of Irrigation and Drainage Systems: A guide for the pEaration of strategies and manuals. With International Commission on Irrigation and Drainage. Paper No. 99. (1989). World Bank Experience with Irrigation DevelopmenL Report No. 8494, Operations Evaluation Departnent. (1990). India: Uttar Pradesh Groundwater Development. Issues and Options. India Agriculture Operations Division. (1991). India: Review of Rainfed Agriculture and Watershed Development India Agriculture Operations Division. (1988). India: Review of Wasteland DevelopmeitL India Agriculture Operations Division. (1988). World Bank wllnternational Commission on Irrigation and Drainage, (1990). Irrigation and Drainage Research. A proposal for an internationally-supported program to enhance research on irrigation and drainage technology in developing countries.