India - Second National Dairy Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 16218 IMPLEMENTATION COMPLETION REPORT INDIA SECOND NATIONAL DAIRY PROJECT (CREDIT 1859-IN/LOAN 2893-IN) JANUARY 14, 1997 Agriculture and Water Operations Division Country Department II South Asia Region 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.  CURRENCY EQUIVALENTS Currency Unit = Indian Rupees (Rs.) Rs. 13.50 = US$1.00 (Appraisal: November 1987) Rs. 22.00 = US$1.00 (Intervening years: 1988-95) Rs. 34.5() = Completion year: 1996 WEIGHTS AND MEASURES 1 hectare (ha) = 2.47 acres (ac) I kilogram (kg) = 2.205 pounds (lb) I kilometer (km) = 0.6214 miles (mi) 1 meter (m) = 3.281 feet (ft) I square kilometer (kM2) = 0.385 square miles (mi2) I liter = 0.264 gallons I metric ton (MT) = 2,205 pounds FISCAL YEAR OF THE BORROWER Apnrl I to March 31 Vice President D Joseph Wood Director R. Drysdale Division Chief/Manager : S. Barghouti Staff Member Herman van Wersch, Pnnc. Operations Officer  FOR OFFICLAL USE ONLY ABBREVIATIONS AND ACRONYMS AHID Animal Husbandry Department (in the states) Al Artificial Insemination AMUL Anand Milk Union Limited CDP Cooperative Development Program CFP Cattle Feed Plant DCS Dairy Cooperative Society EEC European Economic Community ERR Economic Rate of Return EU European Union FAO Food and Agriculture Organisation of the United Nations FMD Foot and Mouth Disease FRR Financial Rate of Return GOI Government of India HF Holstein Fnresian ICB Intemational Competitive Bidding IDA Intemational Development Association [DC Indian Dairy Corporation LCB Local Competitive Bidding LPD Liters Per Day MIS Management Information Systems M&E Monitoring and Evaluation MLPD Million Liters Per Day MPUI Milk Producers Union NCAER National Council of Applied Economic Research NCDFI National Cooperative Dairy Federation of India NDDB National Dairy Development Board NDP-I First National Dairy Project NDP-I1 Second National Dairy Project NMG National Milk Grid OF Operation Flood ROI Retum on investment SAR Staff Appraisal Report SF State Federation SMP Skimmed Milk Powder SNF Solids Non Fat SOE Statement of Expenditure TMDD Technology Mission for Dairy Development UAS Uniform Accounting System UHT Ultra High Temperature UMB Urea Molasses Block This document has a restricted distribution and may be used by recipients onry in the performance of their official duties. Its contents may not otheruise be disclosed iLhout World Bank authorization.   IMPLEMENTATION COMPLETION REPORT INDIA SECOND NATIONAL DAIRY PROJECT (Cr. 1859-IN/Ln. 2893-IN) Table of Contents PREFACE .................. ....................................................i EVALUATION SUMMARY ............................................................. . ...... ii PART I: PROJECT IMPLEMENTATION ASSESSMENT ..........................................................1 A. PROJECT OBJECTIVES ......................................................................1 Statement of Project Objective ...................................................................... 1 Evaluation of Project Objectives ..................2.............................................. 2 B. ACHIEVEMENT OF PROJECT OBJECTIVES .............................3.........................3 Physical Objectives ..................................................................... 4 Financial Situation of Participating Entities ................. ................................. 12 NDDB's Role and Performance ..................................................................... 15 C. IMPLEMENTATION RECORD AND MAJOR FACTORS AFFECTING THE PROJECT ..................................................................... 17 D. PROJECT SUSTAINABILITY ..................................................................... 18 E. BANK PERFORMANCE ..................................................................... 19 Identification and Preparation ..................................................................... 19 Appraisal ..................................................................... 19 Supervision ...................................................................... 21 F. BORROWER PERFORMANCE ..................................................................... 22 Preparation. ..................................................................... 22 Implementation ..................................................................... 22 G. ASSESSMENT OF OUTCOME ......................................... ............................ 24 H. FUTURE OPERATION ..................................................................... 24 I. KEY LESSONS LEARNED .24  PART II: STATISTICAL ANNEXES Table 1 Summary of Assessments .......................................... 27 Table 2 Related Bank Loans/Credits .......................................... 29 Table 3 Project Timetable .......................................... 30 Table 4 Loan/Credit Disbursements: Cumulative Estimated and Actual ............. 31 Table 5 Key Indicators for Project Implementation .......................................... 32 Table 6 Key Indicators for Project Operation . ......................................... 33 Table 7 Studies Included in Project .......................................... 33 Table 8A Project Costs .......................................... 34 Table 8B Project Financing .......................................... 35 Table 9 Economic Costs and Benefits .............. ............................ 35 Table 10 Status of Legal Covenants .......................................... 36 Table 11 Compliance with Operational Manual Statements .................................. 38 Table 12 Bank Resources: Staff Inputs ............... ........................... 39 Table 13 Bank Resources: Missions .......................................... 39 APPENDICES A. Aide-Memoire .......................... 40 B. Financial and Economic Re-Evaluation .......................... 55 C. NDDB's Role and Performance .......................... 58 D. Borrower's Evaluation Report .......................... 66  i IMPLEMENTATION COMPLETION REPORT INDIA SECOND NATIONAL DAIRY PROJECT (Cr. 1859-IN/Ln. 2893-IN) PREFACE This is the Implementation Completion Report (ICR) for the Second National Dairy Project in India, for which Credit 1859-IN in the amount of SDR 121.2 million (US$160 million equivalent) and Loan 2893-IN in the amount of US$200 million were approved on December 15, 1987 and made effective on April 8, 1988. The Credit was closed on December 31, 1994 as originally scheduled. Final disbursement from the Credit, which was fully disbursed, took place on July 28, 1993. The Loan was reduced by two cancellations, US$53 million in September 1995 and a further US$29.9 million in June 1996. It was closed on April 30, 1996, compared with the original closing date of December 31, 1994. Final disbursement took place on June 24, 1996. The ICR was prepared by an FAO/CP mission'/ which visited India in May/June 1996. It was revised by H. van Wersch (Agriculture and Water Operations Division II) of the South Asia Region and reviewed by S. Barghouti (Chief, Agriculture and Water Operations Division II) and K. Uchimura (Project Adviser). The Borrower contributed to the preparation of the ICR through the project evaluation report prepared by the National Dairy Development Board (NDDB), as well as NDDB's comments on the mission's aide-memoire and the draft ICR. The ICR is based on a review of the Staff Appraisal Report and legal documents, supervision reports and project files as well as field investigations and discussions with Bank staff, officials of NDDB, and the participating state dairy federations, milk producers unions and dairy cooperative societies. " Messrs. K. Selvavinayagam (Financial Analyst, Mission leader), J.A. Phelan (Livestock Industry Specialist) and Ms. Frances Sinha (Cooperatives Management Specialist, Consultant).   ii IMPLEMENTATION COMPLETION REPORT INDIA SECOND NATIONAL DAIRY PROJECT (Cr. 1859-IN/Ln. 2893-IN) Evaluation Summary Introduction 1. The Bank has been associated with the Indian dairy sector since 1974 when it approved three state-level projects (Karnataka Dairy Development Project for US$30.0 million, Rajasthan Dairy Development Project for US$27.7 million and Madhya Pradesh Dairy Development Project for US$16.4 million), and the National Dairy Project for US$150.0 million in 1978, followed by the Second National Dairy Project for US$360 million in 1987. Project Objectives 2. The Second National Dairy Project (NDP-II) was designed in support of the third phase of Operation Flood2 (OF III), India's dairy cooperative development program, building on the experience of the three state-level projects and the first national level project. The objectives of the Second National Dairy Project were to promote the establishment of viable cooperative businesses owned and managed by producers, for collecting, processing and marketing milk products in order to expand rural incomes and improve milch animal productivity. These were to be achieved by: (a) expanding and consolidating the integrated cooperative institutional structures; (b) providing the necessary prerequisites for milk productivity enhancement; (c) establishment and expansion of milk collection, processing, storage and distribution activities; and (d) strengthening institutions and training. 3. Loan covenants. Special legal covenants or agreements that were expected to promote the achievement of project objectives were for the Government of India (GOI) to cause participating states to ensure that state federations, unions and societies follow / The beginnings of India's dairy cooperative development are in the Kaira district of Gujarat and the establishment in 1946 of the Anand Milk Union Limited (AMUL) responding to the limited marketing opportunities for traditional milk producers in the face of the marketing practices of a private dairy monopoly. Operation Flood (OF) built on this experience when cooperative dairy development became an agricultural development priority in the national development plan. The first phase (OF I) covered milksheds in ten states during the period 1970-77, followed by OF II in 1978-86 extending the program to cover virtually the entire country.  111 Operation Flood conditionality and for NDDB to ensure that state federations, unions and societies function under by-laws and charters consistent with OF principles. 4. Evaluation of project objectives. The project, based on the sound foundation of improving rural incomes and milch animal productivity through the integrated dairy cooperative structures for the production, collection, processing and marketing of milk and milk products, was important for the Indian dairy sector. Although the design was based on the well-tested, locally developed 'AMUL' model, the appraisal mission was overoptimistic in assessing the rate of change of state governments to the concept of farmer control at various levels of the cooperative structure, though the SAR highlighted this as a major risk. The project was also complex as it involved many states which have their own priorities and commitments, in addition to the scale and spread of project activities which called for a level of high-quality management which many federations/unions found difficult to command. Significant implementation delays or shortfalls in achievements were foreseen at appraisal. A longer-than-normal implementation period was therefore regarded as acceptable if it helped promote institutional changes to ensure the long-term viability of the dairy cooperative sector. Implementation Experience and Results 5. Physical implementation of the project was slower than envisaged due to the delayed start in some of the participating states. Nevertheless, the project, by and large, has achieved its physical, sectoral policy, financial and institutional development objectives in Gujarat, but only partially in other states. Although they were not stated objectives under the project, gender issues and equity have received substantial support, particularly during the later years of project implementation. While the evidence is not systematic and varies considerably across areas, mission findings from field visits and from various sample surveys indicate that dairying has made a significant contribution to the incomes of small, marginal and landless farmers. The dairy societies provide a market outlet which is much fairer to the milk producers than the private vendor although the societies are not (usually) able to displace fully the pattern of informal credit which binds poorer milk producers to private vendors. In addition, women are increasingly enrolled as members (and in some cases in the managing committees as well), in line with state policies and the objectives of the Cooperative Development Program of NDDB. The delicensing of the dairy sector in 1991 had a negative impact on milk procurement by the cooperative sector. As a result of this and other factors, the current estimate of the project's economic rate of return is 21 percent, down 6 percent from the appraisal estimate of 27 percent. Both appraisal and completion estimates are more speculative than usual, because they incorporate projected returns to investments not yet completed. 6. In major states, the dairy cooperative societies (DCS) have been generating surpluses while the performance of many district milk producers unions (MPU) and state federations (SF) is still unsatisfactory, although they have shown improvement in recent years. At project completion, profit-making units have risen to 63 percent, compared to 50 percent at appraisal. As under the predecessor National Dairy Project (NDP- 1), NDDB was  iv the primary implementing agency. Over 70 percent of its staff have degree or post-graduate level qualifications, leaving no doubt as to the professional competence of NDDB staff. The promotion aspect of NDDB's lending activity acts as an essential catalyst for development and not merely as a conduit through which lending resources pass. NDDB in this way has an impact on tangible factors such as proper management practices and appropriate levels of technology. NDDB's continued claim on a significant portion of GOI's scarce development resources calls for a more critical analysis of its operations and in particular of the allocation of resources placed at its disposal. Its overall financial management has improved and is now generally sound. A principal finding is that NDDB should be encouraged to find ways to further improve its operations rather than that it be dismantled. 7. Project sustainability. The economic rate of return (ERR) of the project is sufficiently robust to suggest sustainable economic performance. The major challenge to the sustainability of the dairy cooperatives is their rational development with proper staffing and maintenance of a sound financial structure to ensure solvency and liquidity for commercial survival. This requires that the entities are owned and directed by farmers under some form of participatory decision-making mechanism assuring autonomy in pricing, investment and marketing decisions, efficiency in production and processing, and guaranteeing better incomes to farmers and the lowest possible price to consumers. These conditions exist in Gujarat but only partially in the other states. Consumer prices in metro cities such as Bangalore, Calcutta, Hyderabad, Madras and New Delhi are administered. As these prices are not likely to be increased in the immediate future, the options available to the federations/unions are to introduce cost-saving measures, better management practices and expansion into high-value product lines. These steps must be complemented by measures for enhancing animal productivity, training and technical assistance to cooperative societies, improvements in financial management, and better market promotion. 8. Development of fully sustainable operations in a consumer industry requires the involvement of a regulatory body to assure product quality. Thus far, NDDB has performed this function. However, as the apex institution of the cooperative dairy system, responsibility for quality assurance as well as other industry-related issues constitutes a conflict of interest for NDDB. Government needs to establish an industry body, of which NDDB would be a member, to look after such matters. Operating costs of such a regulatory body would need to be shared between the industry and government. 9. Actual cost, financing, and implementation timetable. At appraisal, it was estimated that the project would be completed by December 31, 1994. Current estimates indicate that final completion is not likely to be before June 30, 1997, although the Credit was closed on April 30, 1996. The final cost of the project is now estimated at Rs12,553 million (US$503 million) compared to the appraisal estimate of Rs9,150 million (US$678 million), representing a 37 percent cost overrun in nominal rupee terms. The Credit was able to finance about 66 percent of total actual expenditure compared to 53 percent of the project costs at appraisal, mainly due to an increase in the disbursement percentage.  v 10. Key factors relating to achievement of project objectives. The most important shortfalls which hindered project implementation can be classified into: Factors not generally subject to Government control. These included employee unrest, though short-lived, in a few unions. Factors generally subject to Government control. These included delays in land acquisition, lack of reliable electricity and water supply, failure or delay in full adoption of OF principles, heavy accumulated losses and poor financial position of SFs and MPUs in some states, and lack of freedom in setting prices in many states. Factors generally subject to implementing agency control. These included delays in appraisal of sub-projects, delays in the implementation of institutional strengthening measures such as monitoring and evaluation (M&E). 11. Performance of the Bank and Borrower. The performance of both the Bank and Borrower has been mixed. Both have satisfactorily carried out project identification and preparation, and appraisal was also satisfactory, overall. Supervision of the project by the Bank received inadequate attention while covenant compliance by the Borrower has only been partial. 12. Project outcome assessment. The project outcome is rated satisfactory. Summary of Findings, Future Operations and Key Lessons Learned 13. Findings. The institutional, managerial and technical adjustments promoted under the project have helped the dairy sector to improve its performance and set the stage for sustained development of a competitive commercially-oriented domestic dairy industry able to respond to increasing domestic demand for milk and milk products. The project's re-estimated economic rate of return is robust. A matter of particular concern is the still poor, though substantially improved, financial performance of many state federations and district unions. Continuing emphasis is required on the commercial and financial viability of cooperatives. Equal emphasis, if not more, is required to sustained measures aimed at enhancing milk yields. NDDB is aware of these shortcomings and has initiated programs to address these issues. Given the quality of its senior management and its competent and loyal staff (many counting over 20 years of service), NDDB's continued attention to these issues is likely to be assured. 14. Plans for completion of OF II objectives and future operations. NDDB has indicated that it will: (a) help completion of all unfinished sub-projects by June 1997, and make available adequate funds for this purpose on a full loan basis at an annual interest rate of 14-15 percent; (b) continue the cooperative development program covering selected  vi DCSs; (c) intensify its efforts to rehabilitate selected state federations and MPUs on the lines adopted in Mysore union; and (d) encourage cooperatives to upgrade their productivity enhancement programs. While the Tamil Nadu federation and unions have utilized their own funds to carry out sub-projects as a result of the embargo on fund release to that state by NDDB, Maharashtra which has sub-projects that are incomplete due to their late start-up is reluctant to accept funding on a loan basis at 14-15 percent p.a. The subject is under discussion between the concerned states and NDDB. 15. Lessons for future projects. The main lessons learnt are: (a) In any project involving farmer-controlled cooperatives, an understanding of the proper role of government in project implementation is crucial. Project design should explicitly take this role into account. The AMUL model, a successful experience in milk collection, processing and marketing, controlled and directed by farmers in Gujarat, was considered feasible for replication in other states which had different cooperative traditions and culture. The model hinges on autonomous farmer control at DCS, union and federation levels. Many other state governments do not fully support the concept of farmer control at the union and federation levels. The requirement of state guarantee for loans taken by unions or federations has not helped the development of a "hands-off' policy of state governments. It would, therefore, appear that a successful expansion of AMUL principles throughout the country requires, in the final analysis, a supportive role by governments to the cooperative movement such as presently enjoyed in Karnataka and Bihar. Accordingly, the bank might have taken a tougher line during project supervision in scrutinizing sub- project appraisals rejecting those of states which did not meet an acceptable level of compliance with the institutional requirements of the project. (b) In a sheltered industry, promotion of competition is essential to enhance operating efficiency. Although private competition has adversely affected cooperative milk procurement in the early years after delicensing, it has helped to raise cooperatives out of their complacency and acted as a spur to efficiency in many cases, forcing them to contract out some of their activities to the private sector like milk transport, packing and maintenance of factory gardens. (c) In project design, especially in the case of complex development programs, particular attention should be paid to assessing the existing situation and to devising ways to overcome or manage existing impediments. In NDP-II, the quality of management was perceived as a principal risk, thus requiring special attention.  vii (d) Development of fully sustainable operations in a consumer industry requires the involvement of a regulatory body to assure product quality. At present this role is performed by NDDB. However, there is a conflict of interest between NDDB's function as the apex body for the dairy cooperative system and an industry-wide regulatory function. Most of the dairy plants visited made heavy demands on NDDB for technical assistance in all aspects of the dairy sector. There was general agreement that the industry would meet some of the costs of a separate regulatory body. However, the Government must also make a contribution to the agency which would be responsible for maintaining quality standards for technically guiding the industry, and for ensuring that good manufacturing practices are applied throughout the dairy industry. (e) Government commitment to resolving sectoral and state-wide issues is critical to the successful project implementation and operation. In NDP-II, solutions to some problems depended on the willingness of state governments to address difficult issues which could not be resolved at project level, e.g. adoption of OF III principles, pricing decisions, amendment to State Cooperative Acts and reduction in staffing. (f) State governments must be fully involved in the design of reforms sponsored by the central government as a basis for project design. Some of the reforms implicit in OF III principles initiated by NDDB covered the true cooperative principles. While it is too early to say whether NDP-II has been successful in having a lasting effect on the dairy cooperative structure, the Bank's positive support to project investment and compliance with OF principles proved valuable to India's drive to increase milk output. (g) For a project of the size and scope of NDP-II, the Bank should have insisted on the establishment and monitoring of key performance indicators aimed at providing intermediate assessments of project outcomes and likely impacts. (h) The Bank should not add new sub-projects toward the end of the project implementation period without first examining both the prospects for timely completion and for achieving satisfactory benefits. The inclusion of additional sub-projects within the last two years before Credit closing (e.g. Maharashtra) has been one of the reasons for the delay in the completion of NDP-II. The experience of NDP-I had already taught this lesson which was not applied to the follow-on NDP-II. While utilization of Credit savings was a major objective of these additions, the prospect of delays in completion and accrual of benefits therefrom should have received adequate scrutiny before this decision was made.  IMPLEMENTATION COMPLETION REPORT INDIA SECOND NATIONAL DAIRY PROJECT (Cr. 1859-IN/Ln. 2893-IN) PART I: PROJECT IMPLEMENTATION ASSESSMENT A. PROJECT OBJECTIVES A. Statement of Project Objectives 1. The Second National Dairy Project (NDP-II) was designed in support of Operation Flood III (OF III), building on the implementation experience of three state-level dairy projects (Karnataka, Madhya Pradesh and Rajasthan, all approved in 1974) and one national-level project (National Dairy Project, approved in 1978) through which the Bank assisted in developing India's livestock subsector. The objectives of NDP-II were to promote the establishment of viable cooperative businesses owned and managed by producers, for collecting, processing and marketing of milk products in order to improve milch animal productivity and expand rural incomes. These were to be achieved by: (a) expanding and consolidating the integrated cooperative institutional structures; (b) providing the necessary prerequisites for productivity enhancement; (c) establishment and expansion of milk collection, processing, storage and distribution activities; and (d) strengthening of institutions, and training. 2. Project cost. The project was to be implemented over seven years from December 1987 to December 1994, and the total project cost including physical and price contingencies was estimated at US$678 million (Rs9,150 million). The IDA Credit of SDR121.2 million (US$160 million equivalent) and Bank Loan of US$200 million, amounting in total to US$360 million (Rs4,860 million) were to finance about 53% of the project costs. The remainder (US$318 rmillion or Rs4,290 million) was to be met out of NDDB resources, which were expected to be supplemented by commodity aid from the European Economic Community (EEC), supplied on a grant basis and equivalent to about US$150 million (Rs2,025 million)." 3. Special legal covenants. Special legal covenants or agreements that were expected to promote the achievement of project objectives were for GOI to cause project states to ensure that state federations, milk producers' unions and dairy cooperative societies follow Operation Flood conditionality and SFs to submit satisfactory annual development plans. In addition, covenants required NDDB to: (a) ensure that DCSs, MIPUs and SFs function under by-laws and Subsequently revised to Rs2,227 M. This understates the value of EC contribution of Rupees 2,227 Million as a minimum, but also has maximum food aid quantities (75,000 tonnes of SMP and 25,000 tonnes of BO), provided that these could be justified as a commodity need.  -2- charters consistent with OF III principles; (b) appraise sub-projects and ensure they meet certain criteria spelled out in the Project Agreement (including a financial rate of return of at least 12%) and institutions receiving NDDB project sub-loans deduct an amount necessary to maintain equity in real terms before paying dividends and bonuses; and (c) take steps satisfactory to the Bank to progressively reduce subsidies. 4. Although the participating states have all agreed to adopt OF principles, despite consistent prodding by NDDB some states have yet to comply in part due to obstructive provisions in the State Cooperative Acts. The enactment of the Andhra Pradesh Mutually-Aided Cooperative Societies Act of 1995 represents a positive new development. It has resulted from the combined efforts of NDDB, the Cooperative Development Foundation and the Cooperative Initiative Project in a state where government interference in the past has manifestly weakened the abilitv of the dairy cooperatives to compete with the growing private sector. It would be highly desirable if other laggard states would follow the Andhra Pradesh example and enact legislation to confer greater functional autonomy on cooperatives, enabling them to frame and amend their own by-laws, conduct elections, hire chartered accountants of their choice, take investment and marketing decisions, and allocate funds independently of the government. However, NDDB, finding it unrealistic to expect the State Cooperative Act to be amended has recommended to the central Government that dairy cooperatives be incorporated under the Cooperative Companies Act. This recommendation is still pending with the Government. Strengthening of sub-project appraisal techniques and procedures was of critical importance for ensuring technically and institutionally sound and financially viable development of the dairy cooperative sector. After following the "return on investment" (ROI) methodology in the initial years with the approval of the Bank, NDDB has from 1993 introduced, at the request of the Bank, revised procedures taking into consideration the main criterion of the financial rate of return (FRR) specified in the Project Agreement. Although action has been taken to charge for services, an increased level of cost recovery has not been achieved by the closing date of the Credit. Evaluation of Project Objectives 5. Claritv. The basic concept of the project - improving rural incomes and milch animal productivity - using the integrated dairy cooperative program for the production, collection, processing and marketing of milk and milk products, was important for development of the Indian livestock sector. 6. Realism. The project design was based on the well-tested, locally developed Anand Milk Union Limited (AMUL) dairy cooperative development model on which GOI decided to build its national dairy development strategy. However, the appraisal mission was overoptimistic in its assessment of the rate of change of state governments to adopt the concept of farmer control at various levels of the cooperative structure, though this was highlighted as major risk area in the SAR. Although the model hinges on autonomous farmer control at DCS, MVPU and SF levels, many state governments do not fully support the concept of farmer control at all levels of the cooperative structure. A successful expansion of AMUL principles throughout the country requires, in the final analysis, a supportive role (without excessive involvement or interference) by state governments to the cooperative movement. 7. Complexity. The project was a massive and complex program in states which each have their own priorities and commitments. The scale and spread of project activities called for  -3- high-quality management and skillful planning of implementation, which many SFs and MPUs have found difficult to command. Insofar as the quality of management at federation and union level was perceived to be the principal risk, special attention should have been devoted to examining the existing constraints and making a detailed investigation about if and how existing impediments could be overcome. Deficiencies such as the poor financial position of SFs and MPUs had haunted the dairy cooperatives in many states in the past, and the pace of change was recognized as a risk. The importance of financial viability of the dairy cooperative structure was explicitly emphasized at appraisal and the Bank had specifically provided for a longer-than-usual project implementation period to help promote institutional adjustments to ensure the long-term sustainability and financial viability of the dairy cooperative sector. 8. Responsiveness to Borrower's circumstances. The project responded well to GOI's strategy to build its national dairy development through the Operation Flood programs. 9. Risks. The willingness of the st.ates to implement OF conditionalities was recognized as the project's main risk. Significant implementation delays or shortfalls in achievements were foreseen at appraisal. A longer-than-usual implementation period (more than seven years) was regarded as acceptable if it helped promote institutional changes to ensure financial viability of the OF dairy cooperative sector. The risk of NDDB not completing sub- projects started under OF III was considered to be small. B. ACHIEVEMENT OF PROJECT OBJECTIVES 10. Overall Achievement. The project, by and large, has achieved its physical, sector policy, financial and institutional development objectives in Gujarat, but only partially in other major states.2 Macro-economic policies were not applicable to the project. Although not stated objectives under the project, gender issues and equity have received much attention under this project and poverty alleviation was highlighted as a project benefit. While the evidence is not systematic and varies considerably across areas, mission findings from field visits and from various sample surveys3 indicate that dairying has made a particularly significant contribution to the incomes of small, marginal and landless farmers. DCSs provide a market which is much fairer to the milk producers than the private traders although the society is not usually able to replace or substitute for the pattern of informal credit which bind poorer milk producers to private vendors. In addition, women are increasingly enrolled as DCS members (and in some cases, in the 2/ Including together with Gujarat. Andhra Pradesh, Karnataka, Maharashtra, Tamil Nadu and Uttar Pradesh, which together account for over 82% of total milk procurement, 76% of total investment costs and 74% of funds released by NDDB. 3/ "Case studies of Successful and Unsuccessful Primary Cooperative Service Society and Milk Producers Cooperative Society in Punjab." J.S. Sidhu and R. S. Sidhu. Indian Journal of Agricultural Economics, July- September 1990; "Performance of Dairy Cooperatives in Saurashtra, and Econometric Analysis." Indian Journal of Agricultural Economics, April- June 1994; and "Present Status and Promise of Dairying in India." R.K. Patel, Indian Journal of Agricultural Economics, January- March 1993.  -4- management committee as well), in line with State policies and the objectives of the Cooperative Development Program (CDP) of NDDB. Physical Objectives 11. General. A comparative statement of targets and achievements of the major project components as on April 30, 1996, is given in Table 5 of Part II. Appraisal (or revised) targets were attained or surpassed with respect to some elements of the productivity enhancement component (DCSs covered by artificial insemination (Al), AIs performed per year, bulls for natural service, bull mother farms, silvipasture, urea molasses blocks (UMB) plants and distribution of mini-kits/varietal demonstrations), the number of milksheds covered, the number of DCSs organized, liquid milk marketing, milk powder storage, drying and chilling capacity and rail milk tankers. Shortfalls were experienced in other elements of the productivity enhancement component such as cattle feed plants, seed processing units, first aid centers, semen production stations and sire evaluation centers, and in milk procurement, rural and metro processing and butter storage. The objectives of the major components and their degree of achievement are discussed below. 12. Cooperative Institutional Structure. After a rapid growth during OF II (1981-86), the dairy cooperative structure underwent a period of consolidation and strengthening under the follow-up program - OF III or NDP-11, as indicated by the key indicators of national dairy cooperative development. Dairy Cooperative Development Indicators Unit Appraisal Actual Percentag targets achievement e Average milk procurement per Mlpd 13.7 11.0 80 year ____ Farm families million 6.7 9.2 137 DCSs '000 70 71.8 103 MPUs No. 190 170 95 State Federations No. 22 22 100 All development indicators show achievement well above appraisal targets except milk procurement and number of MPUs. Liberalization has allowed the entry of the private sector to compete away a share of the milk market particularly through its impact on new plants in certain areas and, as a result, average milk procurement by the DCSs fell short of the targets. The reduced number of MPUs was part of a deliberate strategy of consolidation and strengthening rather than geographic expansion of the unions. To continue sustained progress, the cooperatives - SFs and MPUs - will have to further improve their financial profitability and institutional self- sufficiency.  -5- 13. Productivity Enhancement. A major thrust under OF III was to be an increase in animal productivity supported by measures to promote animal health care, feed and fodder usage, genetic improvement and better animal husbandry. Through the project period, increasing emphasis has been placed on increasing cost recovery for technical input provision. This has been particularly by direct payments by farmers for services used and through a charge on the milk prices The charging policy varies between states and unions and is strongly influenced by the charging policies of the Animal Husbandry Departments (AHD). This is a major issue for the Technology Mission for Dairy Development (TMDD) and the dovetailing of OF and AHD services where progress is still very variable. In particular, extension efforts are very limited. The participatory rural management approach being used on a pilot basis has shown promise in identifying the particular concerns of the farmers but extension should be addressed by TMDD if the productivity enhancement objectives of sustainable and economically viable production are to be achieved in the longer term. 14. The artificial insemination program has been strengthened technically by the policy of developing "cluster" centers serving a number of villages rather than the single centers proposed in the SAR. The Al program is concentrated in a relatively small number of major states and involves both the upgrading of buffaloes and breeding of crossbred cattle. For the cattle, crossing with Holstein Friesian (HIF) blood has become increasingly important and crossbreds with rather high exotic inheritance can be found even in the more resource-poor rural areas - despite the SAR recommendation that expansion of Al with exotic semen be limited to areas with relatively favorable conditions. In some cases (Bihar), a number of farmers have been able to cope surprisingly well with newly introduced crossbred cattle whereas in eastern Andhra Pradesh the introduction of such cattle was generally unsuccessful. The HF appears to be the breed of choice of most farmers, but there may well be a case for the small landowner or landless members to concentrate on the use of Jersey semen. HF is chosen on account of the greater quantity of milk produced, but the Jersey, with its lower fodder requirement and higher butterfat production, could be equally profitable where resources are severely limited. The published reports of the numbers of calves born under the Al program understate the effectiveness of the operation. A particular problem is that there is a substantial trade in pregnant animals of higher genetic potential so that the Al program may have only a small impact on milk production within the breeding village. Nevertheless, there is fairly wide variation in the success rate of Al. Where records are available at DCS level, conceptions to first and second inseminations by Bos cows varies from as much as 73% to as little as 30%. There is a system of re-training inseminators. It is more likely that the effectiveness of the system is variable and depends on the resources which local management are prepared to devote to re-training. Another problem appears to be that many farmers seem to miss observing the subsequent oestrus after AI. Several DCS records show that only 50% of animals show positive when presented for diagnosis three months after insemination. Clearlv, this results in a loss of production through extended calving intervals. Buffalo AI is even more variable, and in the Amritsar area of Punjab many farmers are reverting to natural service. Both these matters merit fiirther attention. 15. In most of the villages visited, the animal health services were given the highest ranking by the farmers in terms of the importance of the input services provided by the unions. The change during OF III from regular veterinary routes to the provision of emergency services and infertility camps appears to have been accepted by the farmers as part of the rationalization  -6- and cost reduction program. Village society secretaries have also been given training in basic animal health care. 16. For livestock feeding, the process of producing urea molasses blocks has been changed but the program is still at a pilot stage. Consideration is being given to the inclusion in the blocks of mineral mixtures and anthelmintic. The greater use of such measures should significantly improve animal health. In India, silvipasture can only play a minor role in feeding improved animal breeds. More recently, the provision of fodder seeds has been given greater priority in addition to the program of distributing mini-kits and varietal demonstrations, but land availability restricts the use of these to the larger landholder. However, landless and smaller landholders can sometimes purchase standing crops from these more affluent farmers. Demonstrations of urea treatment of straw continue to be undertaken but the impact of the program appears to be still limited. If the widespread production of UMBs is resumed, the use of these may well have a much wider rate of adoption than straw treatment. 17. Processing and Marketing Facilities. This component represents the largest share (54%) of the total project cost. Milk collection, processing, storage and distribution were the activities covered under this component. Planned milkshed processing capacity was not fully achieved partly due to incomplete sub-projects in Maharashtra, while drying and chilling capacity was more or less achieved. Liquid milk marketing capacity has exceeded appraisal targets (112%) due to the creation of additional capacities in big cities - Bangalore, Hyderabad and Ahmedabad - other than the originally planned four metro dairies - one each in Bombay, Calcutta, Delhi and Madras. A total additional storage capacity of 44,900 metric tons (MT) was constructed, representing a modest shortfall compared to the appraisal target. At appraisal, it was envisaged that an additional 183 rail tankers would be required for long-distance transportation of milk. However, actual procurement was 169 or 92% of appraised target (97% of revised target). Less storage capacity was required due to reduced reliance on re-combination for the supply of liquid rmilk to major urban centers in the lean seasons. Greater use was made of road transport rather than rail in order to provide a more flexible milk distribution network. Thus, in a sense, failure to meet those targets is a positive indication of flexible management. The performance of the main sub-projects is highlighted in the following paras. 18. Cooperatives are obliged to accept all the milk delivered by producers and the installed processing capacity must therefore be capable of handling peak daily intake. Since there are considerable variations (3:2) in milk production, plant utilization will always theoretically be less than 70% installed capacity. In practice, however, many plants are handling considerably more than rated capacity through operating up to three shifts. The data on rated capacities and actual utilization in the Operation Flood dairies showed considerable inter-and intra-state variation. Gujarat and Maharashtra do better than most. With few exceptions, the units in Madhya Pradesh and Andhra Pradesh operate at low capacity. Some units, however, operate considerably above rated capacity and these are mainly concerned with liquid milk processing where the introduction of three shifts and the additional packaging material has allowed them to process and package more liquid milk. Milk drying and butter making activities have been designed largely to cope with the surplus milk at peak supply, and the capacity utilization is therefore generally much lower and more variable. NDDB has proposed management contracts for some of these cooperatives to upgrade performance and improve competitiveness with the private sector.  -7- 19. The technology and equipment introduced were tried and tested elsewhere and consequently there were no major problems with installations assisted by the project. The NDDB Engineering Group, with the assistance of experts provided by the project, established drawings and specifications for the various sub-projects and offered turnkey consultancy service at a fee - 5% of sub-project cost. 20. Where possible, NDDB prepared standardized specifications and drawings for items of equipment which were common to a number of projects, e.g., storage tanks, pasteurizers, cleaning systems, and also for cattle feed plants. High standards were set for locally produced equipment. However, the expected cost savings were not always fully realized. Nevertheless, capital cost per liter of processing capacity was below Rs500 for most installations of expanded capacity. 21. The financial feasibility calculations were based on high throughput rates for most of the season (>70% utilization capacity) which may be difficult to achieve due to the high priority being given to supplying the liquid milk market. All five major sub-projects initiated in recent years at Gandhinagar, Amul, Kheda, Noida and Calcutta metro dairy involved an increase in the level of automation which was justified as much on the basis of quality and efficiency considerations as on labor savings. The standards for civil works were for a combination of reasons set at a very high level. The first three plants involved a high capacity energy-efficient plant for the production of milk powder, butter and cheese products, which met highest international standards. Drying and mechanized automated cheese plants are capital-intensive and capital cost of Rsl,000 per liter of processing capacity are acceptable, though slightly on the high side. 22. NDDB also played an important role in guiding the evolution of the technology and equipment used within the dairy cooperative sector. There has been a progressive increase in plant capacity and degree of sophistication. It may be argued that some of the most recent plants are too complex, but the impressive level of indigenization of manufacture of the processing equipment and automation would suggest that they are appropriate (for the progressive federations/unions) for the rapidly improving conditions in India. An innovative pilot scheme to substitute bulk chilling tanks for the traditional can collection at village level was successfully introduced in the Kaira district of Gujarat. These centers have potential for better operation of the National Milk Grid, but their replication must be preceded by comprehensive feasibility studies in each location. 23. The Dairy at Noida also has a number of innovative features such as vertical design and an ETP which combines anaerobic and aerobic treatment on a compact site. The total cost of the project is estimated at Rs430 million for 400,000 liters per day (LPD) of pasteurized milk in sachets. The fifth plant at Metro Dairy, Calcutta is impressive by any standards and in addition to 400,000 LPD liquid milk, has facilities for ice cream and UHT milk processing. The total cost is nearly Rs500 million. The cost per liter of processing capacity in both cases is in excess of Rsl,000 and this will constitute an exceptionally high loan servicing cost per liter of milk even at 100% capacity utilization. This must be taken in the context of total operating costs for liquid milk in sachets of approximately Rs1.0 per liter given by both cooperative and private factories supplying urban markets.  -8- 24. Realizing the need for competent management of processing facilities, NDDB has to a large extent, concentrated investment in Gujarat and mother dairies where required expertise was available or secured without much difficulty and trained. 25. Cattle Feed Plants (CFP). Some 11 cattle feed plants have been financed under the project, mostly for expansion of existing capacities and, of these, six units were taken up under NDP-I and completed under NDP-I. The available operating results of all plants indicate that nine completed in 1992/93, eight in 1993/94 and one in 1994/95, have made small net profits. Pricing strategy adopted by the federations/unions for cattle feed involved either making no- profit, no-loss (Gujarat) or selling slightly below cost due to input cost increases ahead of price increase (Karnataka), on the grounds that cattle feed is a necessary input for production of milk within the system and the total operation of the federation/union (and not simply the CFP) should constitute the criterion for financial viability. There is some merit in this argument provided that cross-subsidy by MPUs through DCSs does not penalize farmers purchasing cattle feed from other sources as against those who do buy from DCSs, although it was agreed during negotiations of the NDP-II that investments in CFP should be viable, regardless of other activities of the federation/union. Field visits indicate that some CFP units are incurring losses partly due to substandard product quality which does not help them to achieve profitable levels (Madurai), while some produce good-quality feed, capturing markets in neighboring states (CFP in Punjab and Karnataka). Another reason for low prices could be the desire to attract farmers to the cooperative fold. Overall, action has been taken by NDDB to overcome these problems in several ways including input cost control, range of feed production and feed grid. However, selling below cost is a bad strategy from the society's point of view because efficiency gains from competition by the private sector are stifled. 26. Objectives of Sector Policies. The cooperative sector policies relevant to the project included the adoption of Operation Flood principles, milk pricing and subsidies. These objectives were only partially achieved. As of April 30, 1996, Gujarat continued to be the only state in the country in which ownership and control of the milk industry was fully in the hands of farrmers at all three levels - DCS, MPU and SF. All other states continue to exercise varying degrees of control through appointed administrators at union/federation level, usually the latter. Progress in regularizing the situation is slow, thus delaying the application of the OF principles nation-wide. The lack of compliance by the states in this respect was considered the most serious impediment to OF generally under NDP-I and at appraisal of NDP-II, and persists as the major outstanding issue at the completion in NDP-II. The state governments have controlled both producers' and consumers' prices of milk in some states, seriously affecting the growth and financial viability of the dairy cooperatives. Milksheds where the price is decided by state governments/union territory include Bhavnagar, Kutch, Surendranagar and Junagadh in Gujarat, and Sikkim. Milksheds where price is decided by the federation in consultation with the unions, include all the milksheds of Haryana, Karnataka, Kerala, Orissa and Tamil Nadu (where the state government is also consulted). As regards subsidies, specific steps have been initiated to charge for services provided. However, the level of charges is still too low in many cases to satisfy full cost recovery. 27. Financial Objectives. The project has fully achieved its financial objectives in DC Ss (except those defunct or dormant which have not yet been liquidated), but only partially in SF/MIPUs. The shortfall is largely due to the lack of adoption of OF III principles fully except in  -9- Gujarat, excessive staffing levels in SFs and MPUs, and limited freedom in setting realistic prices. This issue is discussed in detail in paras 39-46 as the financial viability of cooperative businesses was an overarching objective under the project. 28. Institutional Development Objectives. The project impact on institutional development was mixed. The main means of achieving the institutional development objectives were by the provision of infrastructure and equipment, training of NDDB and union level personnel, studies related to marketing and management, monitoring and evaluation activities. These were all provided to varying degrees of success. Staff training has been provided as planned. Various studies have been carried out with grant funding from the EU. As a result, NDDB, which acted as project promoters, consultants and financiers, gained a high level of promotional ability, technical skills, financial expertise and overall control, without which the considerable gains made would not have been possible. Its innovative approach to milk procurement and marketing, and rural organization, its dynamic management, market and coordination activities are all highly commendable. 29. On the other hand, the managerial and financial performance of SFs and MPUs have generally suffered due to a number of reasons: the common cadre system in some states (Andhra Pradesh, Karnataka, Tamil Nadu and Uttar Pradesh), with frequent transfers has impeded proper management, accountability and stability of operations. The information system is largely designed to provide data on milk procurement, marketing: and inputs. In certain respects, the information generated does not provide a valid basis for monitoring results or evaluating impact. For example, in two key areas -- socio-economic impact (or income effects on different socio- economic groups), changes in animal productivity and progress of sub-projects -- there is scope for streamlining and clearly defining the parameters (member-pourers/non-member pourers; number and type of animals as well as members' landholding on a sample basis; and time and cost overruns as well as revised FRR for sub-projects) and relating inputs to outputs (e.g. Al). Greater flexibility is also necessary to reflect changes -- for example, the introduction of by-pass cattle feed, and coverage of DCS by government services under the TMDD. 30. Equity and Gender Objectives. Although equity and gender issues were not explicitly stated objectives of the NDP-II, the project has contributed to poverty alleviation and involvement of women in dairy activities. In the absence of clearly defined M&E data on beneficiary participation, it is not possible to make a precise assessment. The preliminary results of a sample survey" indicate that in the OF villages about 78% of total households own milch animals and nearly 63% are members of DCS. About 87% of member households possess operational land while, in the case of non-member households, it is 79%, showing that land ownership is not a distinguishing criterion among member or non-member households. Distribution of members by different social groups show the mix as: scheduled castes (14%), scheduled tribes (4%) and other backward castes (41%). The project's impact on rural incomes is, thus, considerably more equitable than most crop production programs as a result of the much more even distribution of animals between social groups. Female participation is about 20% of total membership of DCSs, representing an important change in an overwhelmingly male- dominated society. Given the social constraints, women's membership does not, by itself, ' "Impact Evaluation of Operation Flood Programme on Rural Dairy Sector, preliminary observations", NCAER, May 1996.  -10- guarantee active participation. Nonetheless, in many areas, and particularly in some women's DCS, women have gained an increase in mobility, in self assurance and some access to income. 31. Another significant contribution under OF III has been the emerging realization that cooperative development is not just about numbers - numbers of DCS or numbers of registered members. Members may be registered but that does not mean that they are involved in the cooperative, or if they are involved, that they understand what is going on and participate actively. This has emerged as a critical area as the rapid expansion of the dairy cooperative sector throughout the country has involved a more 'top-down', target-oriented, approach in contrast with the initial process of growth of dairy cooperatives in Gujarat in response to local needs and leadership initiative. The Cooperative Development Program of the NDDB represents a very important, innovative, committed and generally gender-sensitive effort for membership education and institution building at both DCS and district union level. The drawbacks lie in the scale of the task involved and the need to achieve a balance between more intensive follow-up (which limits coverage but improves effectiveness) and wider coverage (which may reduce effectiveness). Impact studies of the Cooperative Development Program have not contributed as much as they might have to achieving this balance since they tend to be limited to assessments of short-term effects. 32. Operation Flood has been criticized as a program which draws on substantial international and government funding for a development program which benefits the better-off, larger farmers. These criticisms tend to be vigorously denied. However, the debate tends to be more emotional than rigorous. A careful assessment of the distribution of benefits needs to be made. This is apparently being done by the NCAER. There may be scope for supplementing this with a careful examination in different regions of existing data at selected DCS combined with careful cross-checking and interviews with DCS members. 33. At the same time, it is apparent that, whilst poorer members may be benefiting from DCS membership: (i) it is larger farmers who have the resources to invest in more and improved animals and to feed and maintain them very efficiently; and (ii) marginal farmers and landless families do not have the same level of resources and usually find it more difficult to maintain improved animals efficiently. 34. Project Impact. The institutional, managerial and technical adjustments promoted under the project have helped the dairy cooperative sector to improve efficiency and productivity, and set the stage for sustained development of a competitive commercially-oriented domestic dairy industry capable of responding to the expanding domestic demand for milk and milk products. An outstanding impact of the project is the continued development of milksheds, making them attractive for the entry of the private sector which itself was facilitated by the delicensing of the dairy sector in 1991. In addition, the project can claim substantial credit for the impressive gains made in annual milk procurement and to some extent productivity of milch animals; increase in per capita milk consumption; expansion in number and capacity of dairy plants; creation/expansion of infrastructure for greater procurement, processing, and marketing of milk, produced in rural areas which, in turn, has increased urban milk supplies; increase in the number, membership and earnings of DCSs; evening out seasonal and regional imbalances in milk supply; expansion of dairy equipment manufacture to meet country's demand indigenously; and emergence of the country as a potential exporter.  -1 1- 35. The project had a major impact on rural incomes and employment of a good proportion of small, marginal and landless farmers. It benefited about 4.1 million additional rural households (22.5 million people) and continued support towards improving milk yields and family incomes of the 9.2 million households (50.6 million people) already enrolled under OF. Costs per beneficiary household are low, estimated at Rs3,060 (US$90) per incremental household or Rs1,365 (US$40) in terms of total households. As on March 31, 1996, an average milk procurement of 10.7 million liters per day (Mlpd) (or peak procurement of 13.4 Mlpd) has been achieved, representing Rs33.8 billion (or nearly US$1billion) in cash revenues for rural farmers, and a significant increase in the availability of milk and milk products in urban areas. The presence of DCSs in the market has also provided competition, thereby improving retums to producers in general, regardless of the purchasing agent. It has also generated significant, though unquantified, social benefits to the rural community through provision of funds principally for village communal facilities, school/college buildings, and recreational facilities. 36. Economic Rate of Return. The current economic rate of return re-estimated at completion is 21%. The recalculated ERR is based on actual expenditures incurred up to the closing date as well as the costs to complete ongoing activities until June 1997, and an actual and updated projection of benefits. Underlying assumptions about costs and benefits, and other information supporting the analysis, are shown in Table 9 of Part II and Appendix B. The re- estimated ERR differs from the appraisal estimate of 27%, mainly due to the shortfall in milk procurement. 37. Sensitivity Analysis. The three factors which are expected to influence the ERR are the net value added through rural-urban transfer, the inability of the SFs/MPUs to operate and manage processing plants as designed and, to a lesser extent, the assumed productivity growth. If the net valued added is reduced by 20%, the ERR falls to 19%. Poor operation and management would lower ERR to 5%, assuming that the ineffective operation of SFs/MPUs would depress project benefits by 20%. The decline in value of incremental production by 20% would lower ERR to 9%. Thus, both management and incremental value of production have a critical influence on the ERR. 38. Financial Re-evaluation. Project investments have ranged over a variety of new and expanding processing capacities and feed mills located throughout the country. Appendix B shows the FRR on models estimated at appraisal and at completion, which are summarized below. FRR (%) Appraisal Completion Average DCS 15% 17% New MPU dairy plant 20% ) 30% Existing city dairy plant 28% ) Gandhinagar plant 18% 12% Cattle feed plant 16% 9%  -12- At completion, financial results are more or less similar to those estimated at appraisal except for the Gandhinagar plant and cattle feed investment. Considering the growing diversification of city dairies, a single analysis has been done for city dairy/MPU dairy plant at project completion. The lower results of the Gandhinagar plant are attributed to low capacity utilization while the lower FRR for the cattle feed plant was not unexpected due to the policy of the investing SF/union operating these plants on a no-profit no-loss basis. Financial Situation of Participating Entities 39. Overall Performance. In major states, DCSs have been generating surpluses and building up their liquid balances in the form of deposits with banks or with their unions and investments in approved securities. The financial performance of many SFs/MPUs is still unsatisfactory, although they have shown improvement in recent years. At appraisal, about 50% of participating borrowers (SFs, MPUs and city dairies) were earning annual profits with the other 50% either earning a cash profit (i.e. excluding depreciation and interest accrued but not yet due) or varying levels of losses. At project completion, profit-making units have risen to 63% (85 out of 13 6 entities). Moreover, there is a definite trend of improved performance. In addition to the rnilk operations, several states have cattle feed operations which are run on a no-profit no-loss basis (see para 25). 40. Implementation experience indicates that SFs and MIPUs operating substantially within OF principles (substantial farmer control and direction, and independence from state governments) have the more profitable operations (e.g. Gujarat, Karnataka and Maharashtra). On the other hand, other SFs and MPUs operating with state government intervention (e.g. Andhra Pradesh, Haryana, Tamil Nadu and Uttar Pradesh) usually suffer from inadequate selling prices, overstaffing and less motivated management. The overstaffing is an inevitable consequence of the cooperatives taking over facilities and staff from government without the necessary re-structuring. An exception is the Gujarat Federation which was very careful to avoid doing this. Other important causes for losses have been insufficient development of milk procurement and marketing potential, and limited diversification into high-margin milk products. The SAR provision for equity build-up resulted in increased equity in many unions but not far enough to sustain the high risk of operating losses. 41. Poor management has also affected the financial performance of unions and federations. To an extent this is because the dairy industry has become more competitive and contestable. One of the reasons for this is the recent (1991) delicensing of the dairy industry. Thus, the real question facing the dairy cooperative sector is: is an MPU in a position to develop and institutionalize capabilities to the point of earning attractive returns? This breaks down into two further questions. Can an MPU develop capabilities to increase revenue per unit? Can it develop capabilities to decrease cost per unit? Investment in facilities for diversified products has been a major attempt to increase revenues. Operating costs are variable (transport costs, product losses). The greatest leeway is in capital cost (fixed costs) and selling price, and if MPUs could reduce this cost and increase the selling price to a tolerable level, then their profitability should be possible. 42. None of the societies visited by the ICR mission were able to raise share capital voluntarily from members beyond the minimum share ofRslO per member. As a result, the build-  -13- up of equity was slow until later years of the project when unions made deductions from price differences or bonuses paid to members in order to beef up capital (e.g., Gujarat, Karnataka, Maharashtra). In other states like Tamil Nadu, share capital had remained constant until efforts were made to convert some of balances in reserve funds in DCSs into the capital of the union. 43. . Financial Profitability. Financial profitability of the federations/unions has long been a matter of serious concern even in the previously completed projects and during supervision of the NDP-II carried out by the Bank. For the purpose of analysis under NDP-II, six major states have been selected, based on a combination of milk procurement levels, investment costs and financing for sub-projects, as indicated below. Because of the interdependence of the individual units within the three-tier structure, financial analysis, in substance though not in legal form, must treat them as one integrated whole, particularly because of the impact of transfer prices between the unions and federations. Viewed collectively in this light, financial position of the integrated structure would appear better due to returns made by profitable unions and practically all DCSs. However, as each unit in,the tier is a profit center, analysis of individual units makes sense from a performance evaluation angle. Milk procurement Investment costs Release of funds by NDDB (volume in 'OOOs) (kg per day) (%) Rs million (%) Rs million Gujarat 3,313 29 3.898 36 3,843 40 Uttar Pradesh 819 7 1.201 11 806 8 Karnataka 1,098 10 951 9 860 9 Tamil Nadu 1,203 11 735 7 602 6 Maharashtra 1.908 17 728 7 391 4 Andhra Pradesh 848 8 709 6 628 7 Other states 2,091 18 2.689 24 2,472 26 All India 11I280 100 10 911 100 9,6041 _O0 44. As noted above, financial performance is generally improving. NDDB records indicate that there had been wide variations among the assisted units within a state and between states, with Gujarat and to some extent Maharashtra ranking well above others. The change in the number of profit/loss making unions and federations and their associated value of profit or loss is shown below.  -14- 1992/93 1993/94 1994/95 State Unions & Federations l No. making Profit No. making Profit No. making Profit or or or (Loss) (Loss) (Loss) (Rs M) (Rs M) (Rs M) Net Net Net Net Net Net profit loss profit loss profit loss Gujarat 15 - 57 13 2 28 12 3 36 Uttar Pradesh 10 13 (33) 11 12 (49) 12 11 (3) Karnataka 2 11 (131) 3 10 (125) 10 3 115 Tamil Nadu 5 6 (171) 6 5 53 6 5 (3) Maharashtra 9 2 (7) 10 1 28 10 1 112 Andhra 2 10 (168) 2 10 (121) 6 6 (58) Pradesh As shown above, Andhra Pradesh and Karnataka show significant changes in numbers, while all have improved their profits or reduced their losses by 1994/95. This is confirmed by ratio analysis as indicated below. State Unions Return on Turnover (%) Return on Total Assets (%) & Federations 1992-93 1993-94 1994-95 1992-93 1993-94 1994-95 Gujarat a/ 0.31 0.14 0.16 5.57 1.38 1.18 Uttar Pradesh -1.35 -1.85 -0.10 -12.58 -16.07 -0.80 Karnatakab/ -3.62 -2.79 2.53 -24.81 -20.04 20.29 TamilNaduc/ -3.91 1.01 -0.05 -69.81 20.54 -1.27 Maharashtra d/ -0.12 0.45 1.62 -2.24 7.99 29.12 Andhra Pradesh -4.76 -3.18 -1.50 -37.94 -27.88 -13.57 a/ Taxation on Gujarat federation, Anand (Unit 3) increased substantially from 31% of profits to 80%. Profits after tax therefore fell fromRs3O.5 toRs. 7.9 million: a74%reduction. As aresult, profits after tax fell by 51%between 1992-93 and 1993-94; they recovered somewhat the following year. Tumover has increased in a steady fashion. Total assets doubled from 1992-93 to 1993-94 when the assets of the federation are counted. Between 1993-94 and 1994-95, Kaira (Unit 2) increased its own assets by approximately Rs300 million and total assets saw the same absolute increase. Faced with falling profits after tax and increasing sales and assets, return on sales and assets have fallen. b/ Kamataka: losses in first two years with remarkable tumarounds in last year. Steady growth in tumover and assets. c/ Tamil Nadu: shows a fluctuating performance while Uttar Pradesh shows negative but declining rates of retum. d/ Maharashtra: highly variable profits but steadily increasing tumover and assets. In all cases, tumover has been boosted by the new processing units coming into production combined with increases in selling price of milk.  -15- 45. Efficiency Measures. It is important to get away from the idea that it is possible to encapsulate cooperative performance as a whole in a single figure of profitability. This is because the degree of profitability in dairy cooperatives is determined mainly by the level of selling prices which, in turn, are strongly influenced by the policies of the state governments. Assessment of dairy cooperative performance in terms of profitability may thus reveal more about government policy than about the cooperative's own efficiency. A variety of performance indicators would be necessary to reveal the different aspects of performance. For this purpose, four indicators have been prepared and presented below. Extrapolating these results state-wide, a tentative conclusion is that Gujarat, Karnataka, Maharashtra and Tamil Nadu - show a more efficient performance than the others 46. Past Losses. All the major states except Gujarat and Maharashtra have federations and some unions with negative net worth, as a result of accumulated past losses, with some predating the NDP-II. Most of them need to be recapitalized while the rest will have to be liquidated. The main contributory factors for poor financial performance include administered prices of milk in metro cities like Bangalore, Calcutta, Madras and New Delhi; poor management as in Jalgaon or Dharwar, and low capacity utilization due to competition from the private dairies as in Andhra Pradesh, Punjab and Uttar Pradesh, and overstaffing in many units. Audits pending for periods ranging for some entities from one to seven years have affected the initiation of corrective measures. Unions Indicators Procurement Turnover per Net profit/(loss) Assets per per employee employee per employee employee (litres) (Rs.'000) (Rs. '000) (Rs. '000) Sabarkantha (Gujarat) 328 1,532 4 155 Kanpur (Uttar Pradesh) 143 741 32 104 Mvsore (Karnataka) 321 1,227 33 242 Salem (Tamil Nadu) 497 1,293 8 587 Kolhapur (Maharashtra) 802 4.000 149 211 Krishna (Andhra Pradesh) 49 269 2 27 NDDB's Role and Performance 47. Management. As under NDP-I, NDDB was assigned the overall responsibility for implementing the project. As part of this, the Board has promoted, appraised, financed, and guided federations/unions and their investments. The promotion aspect of NDDB's lending activity supported by project funds and involving 'directed credit' to processing plants makes the essential difference as to its functions as a real catalyst of development and not merely a conduit through which lending resources pass. The Board in this way has an impact on tangible factors such as proper management policies and appropriate levels of technology. Total staff of NDDB  -16- numbers 1, 1 19, of which over 70% (832) have degree or post-graduate level qualifications. There is no reason to doubt the professional competence of the staff to carry out the tasks assigned to them. NDDB's continued access to a significant share of national development resources calls for a more critical analysis of the Board's operations and in particular of the allocation of resources at its disposal. 48. Operating Procedures. India's protected environment, her shortage of capital, abundance of under-employed labor and her need for exports are cause for reviewing the way that sub-projects are chosen and steered. In evaluating such sub-projects, NDDB has used ROI methodology (changed to FRR method since 1993) which relates profit to capital employed. The terms 'capital' and 'profit' are capable of many interpretations and can be expressed at various levels according to purpose of use. The method also ignores the residual value of assets such as land and buildings. The most important defect is that it fails to recognize the phasing of costs and revenues (time value of money) and concentrates on financial flows in accounting terms rather than cash flows. NDDB has also not introduced ERR calculations, even for large projects, in their evaluations. In this, NDDB's procedures lag behind those of the Industrial Credit and Investment Corporation of India. It is not axiomatic that the plans to establish/expand a particular plant are right in terms of national economic advantage. A financial analysis is no substitute for consistent economic appraisal in selecting sub-projects in terms of national advantage. A review of INDDB's operational procedures suggests that the major deficiencies are unwarranted reliance on the state plans and consequently no use of ERR in decision-making process and to some extent, inadequate attention to monitoring and evaluation of investments, which is largely due to lower priority being given to M&E than to lack of systems. 49. Financial Management. Generally, overall financial management seems sound. NDDB's surplus (excess of income over expenses) has risen from Rs150 million to Rs466 million between FY1987 and FY1995 while its total assets have increased over the same period from Rs569 million to Rs27,771 million. Its equity has also increased sharply from Rs150 million to Rs 14,116 million over the same period, reflecting its expanded involvement in oilseeds and more recently salt in addition to its milk portfolio. However, its difficulties are reflected in its outstanding loan portfolio which is affected by arrears. Arrears of interest and principal are equivalent to about 185% of NDDB's surplus. 50. Loan Recovery and Portfolio Quality. Loan overdues have mounted steadily from Rs82 million in 1990/91 to Rs863 million in 1994/95, representing an annual growth rate of 180%. They are still only about 1% of NDDB's loan portfolio. While the rapid growth in disbursements has helped to diminish overdues as a percentage of NDDB's portfolio, they also have masked the true state of portfolio quality. Loan recovery rates have declined steadily from 100% (of principal) and 79% (of interest) to 32% and 36%, respectively, between 1990/91 and 1995/96. Most overdues (over 40%) are over two years past due. Total recovery, thus, might not reflect current performance. Data indicate that NDDB's current recoveries (collections from loans falling due in the current year), although better, are still 43% for principal and 67% for interest, indicating poor quality of its loan portfolio. The deteriorating trend is further evidenced by the growth of overdues which has outpaced the growth of loan portfolio (NDDB's loan portfolio has recorded an annual growth rate of 133% compared to 180% for overdues between 1990/91 and 1994/95). Federations/unions have huge overdue loan and working capital dues, amounting to Rsl,623 million (Rsl,059 million in loans and Rs564 million in working capital  -17- facilities), with Rajasthan accounting for 27% followed by Madhya Pradesh (26%), Andhra Pradesh (19%), Haryana (7%), Punjab (4%) and others (17%). This clearly underlines the deterioration in the quality of NDDB's portfolio and the critical importance to NDDB in achieving a marked improvement in loan recovery rates. NDDB has recognized these problems and undertaken a selective rehabilitation program in Rajasthan, Mysore union, Jalgaon union and currently Dharwar union. A huge effort still remains to be made. 51. Industry Regulation. Development of fully sustainable operations in a consumer industry requires the involvement of a regulatory body to assure product quality. Thus far, NDDB has performed this function. However, as the apex institution of the cooperative dairy system, responsibility for quality assurance as well as other industry-related issues constitutes a conflict of interest for NDDB. Government needs to establish an industry body, of which NDDB would be a member, to look after such matters. Operating costs of such a regulatory body would need to be shared between the industry and' government. 52. An Assessment. These findings concerning the effects of NDDB's financial operations do not appear encouraging for further promotion of the Board's financing activities. However, the main implication of this is to find ways of improving NDDB's operations--not to dismantle it. C. IMPLEMENTATION RECORD AND MAJOR FACTORS AFFECTING THE PROJECT 53. Implementation Record. As in the case of its predecessor, OF III was successfully implemented overall, but it too experienced considerable implementation difficulties. The most common of these was the delay in execution of the agreements between NDDB and the state governments, adoption of the OF principles and cooperative bye-laws, and later introduction of appraisal methodology as specified in the Project Agreement, and delays in allocation of land, adequate water and power by the state governments in some cases. Another factor contributing to implementation delays has been the reorganization and merger of NDDB and IDC which took longer than expected. Despite these delays, the project has made significant achievements, although such achievements in some cases fell short of appraisal targets. The project also demonstrated that technological inputs alone cannot increase milk production if not supported by a year-round market for milk (as provided by the project) at a remunerative price through the "Anand Pattern" milk producers cooperatives. Appraisal targets were surpassed with respect to the number of DCS organized, the number of farm families covered, milk chilling capacity, liquid milk marketing capacity and metro and milkshed dairies. The reasons for the shortfalls in others have been identified and for the most part are being corrected. 54. The main contributory factors for the implementation delays/shortfalls were: - Factors not generally subject to Government control. These include employee unrest, though short-lived, in a few unions. - Factors generally subject to Government control. These included delays in land acquisition (Balaji dairy, Andhra Pradesh), lack of adequate power supply (Dharwar,  -18- Raichur, Bijapur dairy in Karnataka and Kamal dairy in Haryana), lack of adequate water supply (Ongole dairy, Andhra Pradesh, Namakkal, Tamil Nadu) and lack of both adequate power and water supply (Kodakkal, Tamil Nadu), delayed or non- adoption of OF principles in (Andhra Pradesh, Haryana, Madhra Pradesh, Maharashtra, Tamil Nadu and Uttar Pradesh), heavy accumulated losses and poor financial position of SFs and MPUs (Andhra Pradesh, Haryana, Madhya Pradesh, Punjab, Rajasthan and Tamil Nadu), and lack of freedom in setting prices in many states. - Factors generally subject to implementing agency control. These included delays in appraisal of sub-projects, and delays in the implementation of institutional strengthening measures such as M&E. 55. Actual Project Cost. Table 8A of Part II gives a comparative statement of appraisal estimates and final costs. The project, as envisaged at appraisal, was completed by the extended Credit closing date except for some ongoing sub-projects in a few states (e.g. Maharashtra). When allowance is made for costs to complete, the total final cost would amount to RsI2,553 million (or US$503 million), representing a 37% cost overrun in nominal rupee terms, which comprise changes in both quantity and price. Expressed in US dollars reflecting the depreciation of the rupee against the US dollar (and US dollar depreciation against SDR), there would be a cost underrun of 26%. The cost of processing facilities (which constituted 54% of the total cost as against 43% at appraisal) increased in nominal terms by 123%, followed by milk marketing system (89%), productivity enhancement (85%), marketing, M&E, training and studies (44%), technical inputs (-49%), and animal breeding (-72%). It is worth noting that technical inputs and animal breeding costs were influenced by the dovetailing with government services, improved efficiency (e.g. Al centers) and the increased emphasis on input cost recovery. 56. Proiect Financing. Both Credit and Loan disbursements have lagged well behind actual expenditures, before gaining momentum since 1994. The project financed about 66% of total actual expenditure compared to 53% of the project costs at appraisal, mainly due to an increase in the disbursement percentage. The EU provided funding for about 19% of the total actual expenditures and the remaining 15% was met by NDDB. D. PROJECT SUSTAINABILITY 57. The sustainability of the cooperatives, which is in the interests of all participants, requires that they develop rationally, be staffed properly and maintain a sound financial structure to ensure solvency and liquidity for commercial survival. Rational development generating adequate farmer incomes and better milch animal productivity is therefore a key requirement. This means that the three-level units, or two-tiers as appropriate, should only be established where adequate financial strength is present and/or the investment cash flow is supported by the assisted entities and with an adequate return; and where institutional and management aspects are such as to allow a self-reliant (e.g., Gujarat OF cooperatives) or evolution towards a self-reliant entity (e.g., Karnataka), whether it is DCS, MPU or federation. It is also essential that the following conditions are fulfilled: the entities are owned and directed by farmers under some form of participatory decision-making; autonomy in pricing, investment and marketing decisions;  -19- efficiency in production and processing, assuring better incomes to farmers and the lowest possible price to consumers, and generating productivity increase of yield per animal per farm. These conditions, by and large, prevail in Gujarat, but are present in widely varying degrees in other states. Given these circumstances, the major challenge for the sustainability of the cooperative dairy development is to further move the dairy cooperative structure in other states towards adoption of OF principles as practiced in Gujarat. This will ensure provision of an attractive price to retain and expand pouring members. This in turn depends on the ability of the unions and federations to pass on the increased costs in the form of reasonably higher prices to consumers. With the exception of Gujarat, all the other states exercise some control on liquid rmilk prices in the metro cities like Bangalore, Calcutta, Hyderabad, Madras and New Delhi. As the prices in these cities are not likely to be increased in the immediate future, the options available to the unions and federations are to introduce cost-reduction measures and expand into high-value product markets. These steps must be complemented by measures for enhancing animal productivity, training and technical assistance for cooperatives, improvements in financial management, auditing and cost accounting, better market promotion and field monitoring and evaluation. Continuing emphasis is also required on commercial and financial viability of cooperatives. NDDB is aware of these shortcomings and has initiated programs to address these issues. E. BANK PERFORMANCE Identification and Preparation 58. General. The conceptual foundation for the project (increasing milk supply to satisfy the urban demand and improving yield per animal) was based on previous experience under the three state-level projects and one national level project, in line with GOI's development priorities for the livestock subsector. Owing to the rapid expansion of infrastructure during NDP- I and the need to improve profitability of the federations and unions, the emphasis on Operation Flood III was, rightly, placed more on consolidation than on expansion. The Bank's performance in project identification and preparation was satisfactory. Appraisal 59. General. The performance of the Bank during the appraisal process was satisfactory overall, though several shortcomings are noted below. Overall, the project appraisal identified project goals and established concise project objectives consistent with sectoral strategy. Two preappraisal and one appraisal mission with the participation of four, one and seven professional specialists, respectively (including two from the EEC at appraisal), carried out a thorough review of the project concept and components. The missions were fairly well balanced and represented in relevant disciplines -- economics, dairy specialization, agro-industries and marketing, livestock processing and financial analysis. 60. Design. The appraisal mission supported the AMUL model and specifically set out the objective of promoting the OF principles across India based on the observation that, where these principles were applied, cooperatives were successful, in Gujarat and elsewhere. Financial  -20- incentives were expected to draw other states progressively toward adopting the OF principles, backed by the project's supporting conditionalities and the reviews of institutional features and progress of implementing OF principles in the context of sub-project appraisals and the reviews of the Federations' annual reports. The interference of the state governments in the appointment of managing directors of federations, price regulations, etc. proved, however, tobe difficult to resist, resulting in only partial achievement in most states of the OF principles of cooperative management. The setting of project targets for the entire country, instead of by region or individual states, had both negative and positive impacts. This strategy has tended to direct investments to states with high potential for success, such as Gujarat, but it has reduced the rate of expansion of cooperative dairy development in states resistant to the OF principles. On the other hand, by not providing any state with a predetermined share of the project investment, it set the stage for investment levels based on performance and institutional development, helping the shift toward increased financial viability and adoption of OF principles that has occurred during project implementation. Financial viability of the cooperatives was recognized as a critical constraint at appraisal, and sub-project appraisal criteria were fashioned accordingly. There was also a good computerized monthly reporting system of key data for each MIPU and additional semi-annual reports were required from the Federations covering financial status, institutional development and adoption of OF principles, but more searching indicators of financial performance were not developed and would have helped closer monitoring of performance. For example, loan recovery performance, already a casualtv under NDP-I, was not identified for continuous monitoring. Also, inclusion of only the first three sub-projects for ex-ante review by the Bank was too flexible, viewed in the context of development banks in India like the Industrial Development Bank of India which had to submit to the Bank sub-projects exceeding a specified value (for others, a summary statement after approval) for ex-ante review by the Bank. At appraisal, NDDB was expected to require "participating cooperative institutions to appropriate annually from net operating income an amount necessary to maintain equity in real terms before profit distribution to members." This requirement was vague as the level of equity was not defined either in relation to the proposed size of investment by an entity or the riskiness of such investment or the level of debt held by the borrower. 61. Technical. Sophisticated technology of new milk processing plants was directly transferred to other states from the experience of Gujarat. It was too optimistic to expect them all to adopt new technology when the investing SF or MNPU was unable to attract competent personnel. 62. Economic. Appraisal estimates of beneficiaries were exceeded, but cooperative milk procurement was less than anticipated probably due to the loiberalization of the milk market resulting in some market share being taken by the private sector. In the financial analysis, investment needs in incremental working capital and land cost were ignored. Although monitoring and evaluation arrangements were built into project design and also into NDDB's monthly reporting, important variables (such as time and cost overruns, revised rate of return) should also have been included as indicators for monitoring during supervision as a basis for estimating by how much the assessment of the impact of a union or its investment had improved or worsened. 63. Institutional. The SAR identified institutional weaknesses of SFs and MPUs. These were to be tackled through sub-project appraisals, for which guidelines were issued, and  -21- the associated documentation by the Federations on their state of institutional development and adherence to OF principles serving as criteria for cooperative financing decisions. This approach and the supporting covenants provided a stisfactory framework for the progressive strengthening of cooperative institutions during project implementation. Model action plans for addressing anticipated issues as they arose would also have been helpful. 64. Commitment. There was commitment both by GOI and NDDB. The commitment of the participating federations and unions was a recognized risk and changeover to OF priniciples was expected to be progressive and varying in pace by state, spurred on by the project's conditionalities during implementation and by the process of sub-project approval by NDDB. This required a firm stance during supervision of project implementation (paras 66-67), but may also have been optimistic on the expected pace of change given past resistance in a number of states. 65. Skill Mix of the Appraisal Mission. The appraisal team consisted of five Bank professionals and two EEC personnel, representing various specialized skills but without including either a rural sociologist or a management expert. The socio-economic impact of the project as well as managerial problems experienced by the federations and unions would have merited their involvement to identify key constraints and possible solutions as well as appropriate monitoring indicators. Supervision 66. Although supervision missions identified the technical, managerial--and in later years-- the financial problems impeding project implementation, on balance the performance of the supervision missions was deficient. Thirteen missions were fielded over 7.5 years from March 1988 to October 1995, at an average of one every seven months. Continuity in staffing between appraisal and supervision was inadequate, but reasonable between supervision missions. Most missions comprised at least two staff members. A sociologist was not present in any of the supervision missions. Considering the projected impact on poverty and women, as well as emerging management problems in many federations and unions, this represents a gross under- representation of critically useful and relevant expertise, which was ignored even in the composition of the mid-term review mission. 67. Stress on farmers' control of cooperatives including their autonomy in pricing and the need for viable cooperative operations was a continuing theme of the supervision missions. There are other equally important aspects which were given inadequate attention by the supervision missions. The supervision mission of March/April 1989 relaxed sub-project appraisal conditions to speed up the processing backlog, allowing regional NDDB management to make an assessment of milk to be procured and marketing/utilization of procured milk, and examination of investment costs, all to be reasonable. "The resulting recommendation by the Project Finance and Management Services Group to the committee for approval of the sub-projects would include conditions necessary to obtain the commitment of the borrowing cooperative for an in-depth appraisal at a later date, and for the implementation of any resulting recommendations necessary to ensure an efficient and viable operation." This relaxation was in effect a relaxation of Bank standards, which had already been downgraded by the Bank's acceptance of ROI instead of the FRR methodology for the appraisal of the first three sub-projects. This has, in a way, severed the  -22- link between sub-project approval and procurement, as demonstrated by the approval of a sub- project costing Rs1,000 million on the basis of only a FAO technical report, while procurement has been approved without proper financial analysis. The Bank's reactivation of the FRR methodology in 1993 prevented NDDB from proceeding with a number of partially prepared sub- projects. As a result, the Bank was faced with a flood of sub-project proposals submitted just before the original closing date of December 31, 1994. This made things difficult both for NDDB and the Bank. Cooperative projects financed by NDDB required intensive supervision to ensure good standards for sub-project appraisal. 68. The extension of the closing date was another point of contention between the Bank and NDDB. The Bank's performance on this subject was not satisfactory. After agreeing to an expansion of project targets and conveying a positive signal that the Credit would be extended by 15 months, the first extension was given for 12 months until December 31, 1995. The Bank again raised hopes for a further extension when the senior officers of the Resident Mission visited NDDB in 1995. But finally, the Bank agreed only to a further partial extension up to April 30, 1996. 69. The supervision missions have not given a proper performance rating in the initial year --1988. It was strange that the project was rated '1' when delays in audits/final accounts of MIPUs and SFs did not allow verification of ROI methodology. Later supervision reports (December 1989 and October 1990) still gave a rating of' 1' while little or no reference was made to the financial performance of cooperatives (SFs and MIPUs), sub-project appraisals and loan repayment. 70. Physical indicators of project implementation given in the supervision reports (December 1993, December 1994 and October 1995) present a different picture. Reference to incremental achievements, while not relating SAR targets with actuals, will not give directly the degree of achievement, compared to appraisal targets. 71. The Bank anticipated the importance of monitoring and evaluation (M&E), at least in certain aspects of the project and made provisions for this. However, the results of M&E activities did not meet the expectations. 72. Particularly in later supervision missions, the Bank raised issues of concern and showed also flexibility in agreeing to increase the disbursement percentage, ROI methodology for sub-project appraisal and speeding up approval of sub-projects. F. BORROWER PERFORMANCE Preparation 73. General. The project was prepared adequately by NDDB and approved by GOI, as the OF III program to be implemented nation-wide. The OF III program was divided into a number of sub-projects with specific physical and financial targets for each state and by each participating federation and union.  -23- Implementation 74. The direct responsibility of implementation of these sub-projects was entrusted to the respective federations and unions. The NDDB was responsible for appraisal and monitoring of sub-projects as well as rehabilitation plans, supporting centralized functions such as planning, coordination, national milk grid and stabilization, productivity enhancement and support for disease control, bulk procurement of equipment and turnkey assignments. 75. The performance of NDDB and a number of the participating federations and unions was, by and large, satisfactory. During the implementation period, NDDB has assisted the federations and unions in the design and erection of processing plants either on a turnkey basis or on a consultancy basis. Several new measures were introduced to improve performance of the federations and unions. Participatory resource management was initiated in 1993/94 as a pilot project in five milksheds - Kolar in Karnataka, Patna in Bihar, Valsad in Gujarat, Barabanki in Uttar Pradesh and Ropar in Punjab. It helped to enhance milk production through tailor-made technical input package based on the specific needs of the farmers. The Cooperative Development Program, which was aimed at educating the farmers, Boards of Directors and staff of the milk unions, showed encouraging results in terms of increase in membership (including women members), member participation in the managing committees and general body meetings. The quality assurance program implemented in Patna and other dairies contributed significantly to the improvement in quality of processed milk. 76. The success in implementation progress was not without its problems. The rapid expansion of DCS and MPUs under NDP-I would have been difficult to absorb administratively even for the best managed institutions and it has, predictably, placed NDDB's systems, and staff under serious strain. NDDB's record of sub-project approvals suffered. 77. A review of sub-projects approved under the project indicates that 50 units had an ROI between 12-13%, 26 between 14-16% and another 23 exceeding 16%. Thus, about 50% of the sub-projects had a ROI marginally above the cut-off rate (12%), requiring intense monitoring for time and cost-overruns as well as product prices, in order to be able to take timely corrective action. Field visits indicate that identification of the critical factors has never been a problem. However, enforcement of corrective measures (e.g., increase in selling prices) continued to be a stumbling block, given the political realities. The achievement of rates of returns estimated at appraisal became difficult due to lack of enforcement of either the required price levels or staff reduction. 78. There was also increasing concern over the deteriorating loan recovery performance, which required more attention to loan collection than loan disbursements. As a result, a comprehensive state-wide rehabilitation plan is being implemented in Rajasthan and selective interventions are being made in other places - Mysore union in Karnataka, Jalgoan union in Maharashtra and Dharwar union in Karnataka. The implementation of the Integrated Business Plan in Mysore union involving the participation of operational staff was a commendable achievement. 79. Because of non-compliance with OF conditionality, funds were not released to Tamil Nadu. The Managing Director of the Tamil Nadu federation has expressed his dismay at not getting any release of funds for work carried out by the federationlunions. Despite the  -24- embargo, own funds have been utilized to avoid cost overruns and hence to safeguard the viability of the sub-projects. As a result, unions are short of funds to make milk payments on time. As no embargo was applied on turnkey projects carried out by NDDB itself, non-release of funds for work carried by the federation/unions would appear strange, particularly when the state government has promised to hold elections (for the cooperatives) in three to four months. During the ICR mission's discussions with the Minister and the Secretary for Dairy Development, it became clear that such elections would be held and that milk prices in Madras city would be adjusted upwards. Given this situation, it seems prudent to explore funding sources and terms to reimburse claims from Tamil Nadu and from Maharashtra as a matter of priority. This will help to minimize payment delays to farmers. 80. The contribution of the NDP-II project to improving the capability of NDDB is hard to isolate since the Board has gained from concurrent implementation of grant funding by the EU as well as interaction with Bank staff since 1981 and expatriate consultants provided under FAO funds-in-trust arrangement and EU. The major institutional benefits of this interaction include: (a) diversification into high-value milk products; (b) growth in the numbers of women members of the DCS; (c) increased choice of brands for consumers of ghee, butter or milk powder; and (d) quality improvement of processing plants. 81. As a part of a successful national program, the project had the full comrnmitment of GOI and NDDB. Compliance with covenants was, however, partial (see para 4 and Table 10 of Part II). G. ASSESSMENT OF OUTCOME 82. The project's outcome is rated satisfactorv on the basis that the provision of a reliable market for milk is a necessary condition for increased production and productivity. The presence of DCSs has generally forced the traditional traders to pay competitive prices. The project has created viable businesses and these have contributed substantially to the expansion of rural incomes and improved milch animal productivity. NDDB is taking, and will continue to take, action to improve the performance of some of the weaker/ailing units, but optimization of producer benefits cannot be sustained without farmer control and direction of the cooperatives. H. FUTURE OPERATION 83. Following discussions with NDDB on an operations/action plan for the completion of all ongoing sub-projects and also to maximize net benefits, NDDB has indicated that it would: (a) help completion of all unfinished sub-projects by June 1997, and make available adequate funds for this purpose on a full-loan basis at an annual interest rate of 14-15%; (b) continue the cooperative development program covering selected DCSs; (c) intensify its efforts to rehabilitate selected federations and MPUs on the lines adopted in Mysore union (now extended to Dharwar union) or in Jalgaon union; and (d) encourage cooperatives to upgrade their productivity enhancement programs. While Tamil Nadu federation and unions have utilized their own funds due to an embargo on funds release by NDDB, Maharashtra has ongoing sub-projects due to their  -25- under discussion between the concerned states and NDDB. 84. An evaluation of the impact of the Bank's support to the Indian dairy sector from its beginning is currently underway. It is undertaken by the Bank's Operations Evaluation Department and will incorporate findings about the impact of NDP-II. I. KEY LESSONS LEARNED 85. The main lessons learned during project implementation are as follows: (a) In any project involving farmer-controlled cooperatives, an understanding of the proper role of government in project implementation is crucial. Project design should explicitly take this role into account. The AMUL model, a successful experience in milk collection, processing and marketing, controlled and directed by farmers in Guj'arat, was considered feasible for replication in other states which had different cooperative traditions and culture. The model hinges on autonomous farmer control at DCS, union and federation levels. The mission finds that many other state governments do not ftully support the concept of farmer control at the union and federation levels. The requirement of state guarantee for loans taken by unions or federations has not helped the development of a 'hands-off'policy of state governments. It would, therefore, appear that a successful expansion of AMUL principles throughout the country requires, in the final analysis, a supportive role by governments to the cooperative movement such as presently enjoyed in Karnataka and Bihar. Accordingly, the Bank might have taken a tougher line during project supervision in scrutinizing sub-project appraisals rejecting those of states which did not meet an acceptable level of compliance with the institutional requirements of the project. (b) In a sheltered industry, promotion of competition is essential to enhance operating efficiency. Although private competition has adversely affected cooperative milk procurement in the early years after delicensing, it has helped to raise cooperatives out of their complacency and acted as a spur to efficiency in many cases, forcing them to contract out some of their activities to the private sector such as milk transport, packing and maintenance of factory gardens. (c) In project design, especially in the case of complex development programs, particular attention should be paid to assessing the existing situation and to devising ways to overcome or manage existing impediments. In NDP-II, the quality of management was perceived as a principal risk, thus requiring special attention.  -26- (d) Development of fully sustainable operations in a consumer industry requires the involvement of a regulatory body to assure product quality. At present this role is performed by NDDB. However, there is a conflict of interest between NDDB's function as the apex body for the dairy cooperative system and an industry-wide regulatory function. Most of the dairy plants visited made heavy demands on NDDB for technical assistance in all aspects of the dairy sector. There was general agreement that the industrji would meet some of the costs of a separate regulatory body. However, the Government must also make a contribution to the agency which would be responsible for maintaining quality standards for technically guiding the industry, and for ensuring that good manufacturing practices are applied throughout the dairy industry. (e) Government commitment to resolving sectoral and state-wide issues is critical to the successful project implementation and operation. In NDP-II, solutions to some problems depended on the willingness of state governments to address difficult issues which could not be resolved at project level, e.g. adoption of OF III principles, pricing decisions, amendment to State Cooperative Acts and reduction in staffing. (f) State governments must be fully involved in the design of reforms sponsored by the central government as a basis for project design. Some of the reforms implicit in OF III principles initiated by NDDB covered the true cooperative principles. While it is too early to say whether NDP-II has been successful in having a lasting effect on the dairy cooperative structure, the Bank's positive support to project investment and compliance with OF principles proved valuable to India's drive to increase milk output. (g) For a project of the size and scope of NDP-II, the Bank should have insisted on the establishment and monitoring of key performance indicators aimed at providing intermediate assessments of project outcomes and likely impacts. (h) The Bank should not add new sub-projects toward the end of the project implementation period without first examining both the prospects for timely completion and for achieving satisfactory benefits. The inclusion of additional sub-projects within the last two years before Credit closing (e.g. Maharashtra) has been one of the reasons for the delay in the completion of NDP-II. The experience of NDP-I had already taught this lesson which was not applied to the follow-on NDP-II. While utilization of Credit savings was a major objective of these additions, the prospect of delays in completion and accrual of benefits therefrom should have received adequate scrutiny before this decision was made.  -27- PART II. STATISTICAL ANNEXES Table 1: Summary of Assessments A. Achievement of objectives Substantial Partial Negligible Not Applicable Macro policies Sector policies Financial objectives 7 z Institutional development F1 0