Report No. 14564-CE Sri Lanka Nonplantation Crop Sector Policy Alternatives March 20, 1996 Agriculture and Natural Resources Division Country Department I South Asia Region ke~~2 : z-0''0 'f'"''';''.''X'''''''ze''i'WSv.. X t ' '''; ,- ,t,e,~~~~~~~~~~~~< ''t,,.t- ~~~~~~~~, ;E7 j.#t6 *''i,'atF .. 7W.. ;,7 .j !_ ,. ,' ,:. ',,.:'''_ D~ocument of -the' Wordd CURRENCY EQUIVALENTS (December 1995) Currency Unit = Sri Lankan Rupee (SL Rs.) SL Rs.53.98 = US$1.0 SL Rs.1.0 = US$0.019 WEIGHTS AND MEASURES I hectare (ha) = 2.47 acres ABBREVIATIONS AND ACRONYMS ARP - Agricultural Research Policy ARTI - Agrarian Research and Training Institute BOC - Bank of Ceylon CARl -Central Agricultural Research Institute CARP - Council on Agricultural Research Policy CBC - Commercial Bank of Ceylon CBSL - Central Bank of Sri Lanka CRBs - Cooperative Rural Banks CWE - Cooperative Wholesale Establishment DRC - Domestic Resource Cost EPC - Effective Protection Coefficient FCD - Food Commissioner's Department FPS - Floor Price Scheme GOSL - Government of Sri Lanka GPS - Guaranteed Price Scheme HFC - High-value Food Crop HNB - Hatton National Bank HORD - Horticultural Research and Development Institute IIMI - International Irrigation Management Institute LDO - Land Development Ordinance MALF - Ministry of Agriculture Lands and Forestry MASL - Mahaweli Autlhority of Sri Lanka MECs - Minor Export Crops MPCs - Multipurpose Cooperatives NAFNS - National Agriculture, Food and Nutrition Strategy NCRCS - New Comprehensive Rural Credit Scheme OFCs - Other Food Crops, a.k.a. Other Field Crops PB - People's Bank PMB - Paddy Marketing Board RCD - Rural Credit Department (of CBSI.) RFI - Rural Financial Intermediary RRDBs - Regional Rural Development Banks RRDI - Rice Research and Development Institute SANASA - see TCCS SAEP - Second Agricultural Extension Project SSM - Sarvadava Shramadana Movement TCCS - Thrift and Credit Cooperative Society TEWA - Termination of Employees Act WP# - Working Paper No. (6 in total, available on request) SRI LANKA NONPLANTATION CROP SECTOR POLICY ALTERNATIVES EXECUTIVE SUMMARY .................................... i I. THE PROBLEM .................................... 1 Sectoral Goals and Strategies ............................. 1 The Policy Dilenmma ................................... 3 II. MARKETS ........................................ 3 Output Markets ...................................... 3 Rice ........................................ 3 Wheat .4 Other Food Crops (OFC) ............................ 5 Fruits and Vegetables .............................. 6 Input Markets ....................................... 7 Fertilizer . ......................................7 Seeds ......................................... 7 III. INCENTIVE POLICY. 8 Price, Trade and Marketing Policy. 8 Incentive Coefficients - NPCs, EPCs and ESCs. 9 Agriculture versus Industry .11 Agriculture versus Agriculture . 11 Comparative Advantage - DRCs .13 Irrigation Subsidy - Cost and Distribution .13 Irrigation Policy Reform .14 Trade Policy Refonn ................................. 16 Price Fluctuations and Trade .18 This report is based on the output of two workshops in Sri Lanka, four background papers written by local Sri Lankan consultants, and the findings of a mission to Sri Lanka in November-December 1994. Mission members were Robert Hunt and Douglas Lister (Joint Task Managers SAIAN). Shahid Khandker (PSP), Terrence Abeysekera (SAISL), and Peter Bloch, Antonio Brandao, Paul Harrison and Forhad Shilpi (SAlAN Consultants). Input was also received from Peer Reviewers and Departmental management and staff. The Country Director is Ms. Mieko Nishimizu (SAIDR) and the Division Chief is Mr. Ridwan Ali (SAIAN). The Green Cover draft was widely disseminated among Government officials and various institutions and individuals in Sri Lanka in September-October 1995. A revised draft report, together with the drafts of dissemination proceedings and responses to major comments was resubmitted for comment in January 1996. Thereafter, some factual errors were corrected and issues relating to sequencing were included. Background papers were prepared in Sri Lanka by staff and consultants of ARTI (Marketing), CARP (Technology Transfer), IMI (Paddy Land Alternatives), MARGA (Exports), and two independent consultants Messrs. S. Berugoda and R.D. Wanigaratne (Land). Copies are available from the institutions and individuals. IV. RURAL FINANCE .......... ........................ 18 Overview ....................................... 18 Market Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Credit Financing ................. ... ... ... ... ... ... . 19 Interest Rates, Savings and Input Financing ....... . . . . . . . . . . . . 20 Credit Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Credit and Subsistence Agriculture ........ . .. . . . . . . . . . . .. . . 23 Rural Credit System - Outlook ......... . .. . . .. . . . .. . . .. . . 24, Policy Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 V. LAND ........................ ..................... . 25 Land Tenure Policy .............. ... .. ... .. .. ... .. .. . 25 Farm Size ....................... .... .... ..... .. . 25 Land Ownership ................. ... ... ... ... ... ... . 26 Land Utilization and Cropping Intensity ....... . . . . . . . . . . . . . . . 27 Cropping Pattern ................ ... ... .... .. ... .... . 27 Land Market ....................... .... .... ..... .. . 29 Policy Conclusions ............... ... ... ... .. ... ... .. . 29 VI. FARM FAMILY INCOME AND OUTPUT ADJUSTMENT ..... . . . 30 VII. TECHNOLOGY GENERATION AND TRANSFER ...... . . . . . . . . 33 Overview ....................................... . 33 Research/Extension Link ........... .. .. .. .. .. . .. .. .. .. . 33 Policy Conclusions and Proposed Actions ....... . . . . . . . . . . . . . . 33 VIII. PUTTING IT ALL TOGETHER ......... . . . .. . . . . .. . . . . . . 34 The Current Situation - What We have Learned ...... . . . . . . . . . . . 34 Policy Choices - Future Directions for the Sector ...... . . . . . . . . . . 34 The Growth Challenge ............ .. .. .. ... . .. .. ... . . . 35 Reforn Priorities ................. ... .. .... ... .. .... . 36 Sequencing Policy Reform .............................. 37 Government's Stance on Sectoral Reform ..................... 38 Discussion and Dissemination in Sri Lanka .................... 39' BACKGROUND WORKING PAPERS (Separate volume available on request) 1. Domestic Marketing in the Nonplantation Crops Sector of Sri Lanka 2. Policy Incentive, Diversification and Comparative Advantage of Nonplantation Crops in Sri Lanka 3. Outward Orientation, Income Distribution and Employment in Sri Lanka's Nonplantation Crop Sector 4. Rural Finance and Sectoral Performance in Sri Lanka 5. Land Tenure Issues in the Nonplantation Crop Sector of Sri Lanka 6. Working Paper Tables EXECUTIVE SUMMARY Rural Staynation - Resolvinp the Dilemma 1. Growth from nonplantation agriculture remains stagnant and income low. The primary source of this growth has been Other Food Crops (OFCs - chilies, onions, potatoes, etc). Assuming no data errors, OFC output is growing in intensity, as the total area under these crops has not grown. More rapid output growth of these higher valued OFCs could be a way out of the stagnation in the sector, but so far this has not happened. With over 60% of Sri Lanka's population rurally based and mainly in the nonplantation sector, the sector's stagnation is of growing concern to Government, particularly in terms of its implications for rural poverty. Despite its low returns per acre relative to other food crops and Government support for more diversified crop-land use, low-valued paddy remains the dominant crop. Why this is so has been a dilemma. 2. Two rural markets, the off-farm labor market and the land market, are at the root of the output puzzle. Off-farm wage rates are higher than returns to family labor for all crops except paddy. With free water, farm families maximize income through growing paddy and working off-farm. Paddy's relatively low labor requirement maximizes off-farm employment per farm family, while still fully utilizing the land. The lack of marketable title perpetuates this stagnation in agriculture. Unavailability of title constrains farmers from increasing farm size and developing a diversified portfolio of income-generating crops. If the land market was working efficiently, able farmers would increase farm size and farming activities and make the most of the commercial opportunities available to diversify a part of their cropping activities and so increase their incomes. Those with more advantage in off-farm activities would phase out of farming, specialize in off-farm activities and by doing so increase their incomes too. For now, all must do both. The Study 3. Workshops and discussions in 1993/94 in Colombo and background topic papers clearly established the need to explain why more widespread use of diversification as a farm-income increasing strategy is not seen and to draw policy conclusions from the results. Examination of farm input and output markets and technology transfer shows no constraints. Price and trade policies, and rural credit markets are also not responsible. But land and labor markets are seen as major determinants of farmers' behavior. The study's policy recommendations are aimed at getting nonplantation agriculture moving again, assuming that sustainable growth with equity is the desired policy goal. Importantly, during the study it became abundantly clear that access to the off-farm labor market was far more important than farm size for creating equity in income-earning opportunities. Appreciation of the major role played by the off-farm labor market is key to understanding of farmers' behavior and provides the basis for the direction of the recommendations. The Findines 4. Markets Work Well - generally. An important general finding is that, except for land, agricultural markets in Sri Lanka are quite competitive. The ease with which buyers and sellers can participate ensures that these markets work so well. This is very important, as it says that the often talked about "mudalali", or monopolistic middleman, is not really a force of any importance in the sector. 5. The Government and the Markets. Government itself has a major direct presence in most agricultural markets. Being caught between the interests of the farmer and the consumer, Government's own performance here has been erratic, at times taking the side of the farmer and then the consumer. The frequency and unpredictability of changes in these market interventions by Government has heightened market volatility, causing unforeseen losses for producers and traders alike. This has greatly reduced private sector investment in the marketing process, particularly storage, and has increased domestic marketing losses. 6. Towards a development partnership. Instead of reacting to every fluctuation, however minor, Government must trust private sector marketing processes and adjust its own role to a background one of facilitating and promote competition in these processes. To fully convince the markets of the seriousness of its commitment to this approach, Government will need to divest itself of all marketing activities, through selling these off or closing them down. 7. Paddy Costs versus Competitiveness -- > Large Subsidies to Irrigation and Paddy. Overall, Sri Lanka currently shows no comparative advantage in production of rice or OFCs in either major or minor, irrigated or rainfed agriculture. Domestic production is much more costly than imports. Obviously, with appropriate amendment in the incentive structure for farmers and improvement in productivity, this will change. While not all rice production is noncompetitive with imports, a large amount is. This non-competitive output is currently maintained by massive- transfers amounting to US$250 million, or 3% of GDP annually. With no cost recovery, the annual equivalent of the subsidy to major new irrigation systems is as much as Rs.50,000 (US$1,000) per ha per year. This is a transfer from ii taxpayers to farmers, except that, instead of being in cash, Government predetermines how farmers should spend it, i.e. on irrigation. In return, 'armers grow a low-value crop, paddy, giving Govemment and the economy a very low return on irrigation investment. The subsidy to paddy, paid by consumers through paying 30% more than world prices for rice, costs about US$125 million per year. In sum, close to 5% of GDP goes to supporting noncompetitive rice production. The concentration of irrigation expenditure on construction of large scale irrigation systems results in a distribution of the irrigation subsidy heavily skewed to few lucky farmers, about 3 % of all farmers. The situation calls for substantive reform in both irrigation and trade policy. 8. Agriculture versus Industry -> a pro-irrigated agriculture bias. Government's trade policy, anti- agriculrure in the past, caused a bias in private investment away from agriculture. Now, progressive reduction in protection for manufacturing industry, combined with continued protection for import-competing crops (rice, chilies onions and potatoes) and massive subsidies to agriculture (primarily to large scale irrigation schemes) with no cost recovery, has leveled the playing field overall, so that there appears to be almost no difference in protection given to agriculture and industry. But the average situation hides major variations within agriculture. Plantation crops receive very little protection. The main beneficiary from irrigation subsidies is output from large-scale irfigated agriculture, especially the newer schemes. By contrast, production based on minor irrigated agriculture benefits less, and rainfed agriculture not at all. In addition, trade policy provides import competing crops, like paddy and chilies etc, with substantial protection over most manufacturing industry, especially where grown on large scale irrigation systems. 9. Agriculture versus Agriculture --> a bias against export crops. Within agriculture, trade policy's substantial protection against import competition causes a strong bias against export crops. Faced with such treatment, Government's stated policy and exhortation for increasing exports from agriculture has had no real impact on producers. Trade reform is badly needed to level the playing field between import and export crops and between producers and consumers, especially low income consumers. 10. Irrigation's Future - Subsidies or Reform? Sri Lanka's high level of rice self-sufficiency has been very costly to achieve, as has its efforts to alleviate poverty through large-scale irrigation developments. The goals and strategies here clearly deserve to be reconsidered, even if only because of their likely financial unsustainability. To start, it must be accepted that, in economic terms irrigation water, unlike drainage, is a private good, and so should be directed by market forces. For this, the basic legal and institutional structures needed to permit market forces to operate must be developed. These center around enforceable water property rights for individual farmers and water markets. Only this will permit efficiency and sustainability in utilization of this vital resource. 11. Trade Reform - good for Growth and Poverty Alleviation. It is important for Government to come to grips with trade reform right away. Although a phased reduction in import tariffs on rice and other import-competing crops will lower onfarm earnings, this will impact on large farmers only. If reasonably phased, this will be small and well within annual price fluctuations, making it invisible. At the same time, farmers will have the opportunity to phase into production of higher valued export crops. With a decline in rice prices, consumers will benefit, as will most small farmers because they consume more rice than they produce. Proportionately, this will be very important for the poor, both rural and urban. Price stability will improve, as seasonal price fluctuations in world rice markets are far less than in domestic markets. 12. Rural Credit Markets. Presently demand for rural credit is limited because input financing is now largely from "own sources", made possible by the high volume of off-farm earnings. Rural credit markets, though presently limited, are quite competitive and becoming more diverse. Stimulated by the scope for reducing of the cost of credit, new forms of rural banking, via semi-formal, NGO-type, institutions, are developing to meet customers needs and market niches. The lion's share of this business is to meet rural demand for nonagricultural loans, since, as noted, most agricultural needs are currently more than met by off-farm earnings. 13. Concessional credit from the Government, NCRCS loans, are responsible for only a very small share (7.5 %) of input financing in nonplantation crop agriculture. The explanation for the low utilization rate NCRCS loans lies in the relative interest rates for saving and borrowing and the small volume of inputs per farmer. The savings deposit rate is so well below the NCRCS concessional loan rate that it makes sense for small farmers to pay for inputs out of surplus earnings. In the meantime, rural savings are so large relative to purchased inputs, that farmers could well afford to purchase more inputs without credit. This also means that Government's concessional credit program has no impact on current sector performance. Also, NCRCS loans generally do not reach the poor farmers. Yet subsidized interest* rates and losses underwritten in the NCRCS program combine to cost the budget about US$4 million per year. So the program should be discontinued. As for the poorest farmers, the Government may consider a tightly targeted program. iii 14. But all is not well in rural credit markets. With rapid sectoral growth, the demand for rural credit would increase and the shallowness of the rural financial markets would become a problem, particularly in supplying long-term credit so necessary for fixed investment. Furthermore, the poor health of the credit institutions could result in rapid deterioration in loan portfolios and institutional viability in the absence of good management and tight supervision by banking authorities. Also, at present, Government borrowing to meet a large fiscal deficit has pushed up interest rates to a level that adversely affects viable economic activities in all sectors including agriculture. 15. Land - the forgotten factor market. Lack of a land market seriously constrains long-run sector performance. Government land is for lease only; intergenerational transfers are permitted, but not sale or sublease. In the absence of a land market, irrigation scheme settlers wishing to exit agriculture face a choice of abandoning a valuable farm or entering the uncertain informal market. Private sector land sales, the only ones permitted by law, are very few, as disputed title makes these very difficult. Records show an extremely low turnover rate of private land. 16. Land Tenure Policy --> depresses returns to land. Government policy, aimed at equitable distribution of land, has fostered a smallholder agriculture. Land legislation and settlement administration has successively limited farm size and constrained the rights of both owners and operators. Government owns 55 % of nonplantation crop land, mainly in large scale irrigation systems. Private sector land is mainly in older, village-based minor irrigation schemes. Much private land is subject to ownership disputes. Population pressure and societal inheritance practices have caused substantial fragmentation on both private and public land, leaving average farm-size well below the public sector irrigation scheme norm of I ha. Overall, 72 % of farms are below I ha. The cropping pattern is heavily paddy oriented, with an 80% share, despite the relatively low net income per ha. Farmers' inability to increase farm size and so earn an income from farming that is competitive with other possibilities causes them to grow paddy and work off farm. This depresses returns to land and any investments in it, particularly investment in irrigation, and to human capital more productively employed off-farm. 17. Why Paddy? - because off-farm wages are high. Why do small farmers choose to continue to grow a low valued commodity, rice, when higher income crops are available? Simply because, for all crops except paddy, off- farm work pays higher wages than the hourly return to family labor. Faced with the inability to sell land and leave agriculture or expand farm size and specialize in agriculture, the combination of paddy production and off-farm employment provides the greatest income-earning opportunity. Labor Force Survey data, confirmed by individual village surveys, show that, on average, Sri Lankan farmers spend considerable time working off-farm, averaging 40% of family income from this. This pattern of use gives the most returns to farm family labor and so makes sense for the family. Although greater production of traditional import competing crops (e.g. chilies or onions) would substantially expand farm income, because the resulting implicit family wage rate declines to below the off-farm wage rate, this would depress overall family income and so does not happen. 18. Breaking the Land Deadlock! Only Government can provide the means to break the deadlock here and allow the positive elements of market forces to go.to work. For public sector land, full privatization, i.e. fully marketable, freehold title would be the best solution. In the absence of full privatization, long-term leases with provision for transfer would be a major improvement. Legislation and programs to establish public and private sector institutions and professions to facilitate development of land markets are also required. Only in this way can farm size adjust automatically to changing commodity prices and increased risks. For private sector land, incentive systems to promote rapid resolution of disputed titles (i.e. tax penalties and rewards) are an immediate necessity. Land values in irrigation settlements depend heavily on irrigation cost recovery policy. Therefore, amendment of irrigation cost recovery policy should be transparent, and phased to avoid overvaluation and subsequent financial stress in land markets. 19. Subsistence Farmers - a problem within a problem. Where farm size is too small to generate an adequate income and the family has insufficient access to the off-farm labor market, a problem of noncommercial or "subsistence" agriculture arises. Trapped in terms both of farm size and employability, these are non-commercial or subsistence farmers. Broadbased intervention policies, such as floor prices or subsidized inputs, transfer financial benefits to commercial agriculture, but have little impact on subsistence farmers' small incomes from farming. Separate programs need to be developed to assist here But to succeed, program design must reflect a full appreciation of the current situation, especially the obstacles to growth for both subsistence farmers and the sector overall. 20. Extension and Research - a circular trap? Agricultural extension, the means for technology transfer, suffers from an "identity crisis". This is due to the supply driven nature of the present service in a demand drivent world. Organized along commodity lines, it is largely focussed on physical output, while its customers are concerned with income. Periodically, it restructures itself in an attempt to come to grips with reality, but its inadequate incentive iv structure leaves these efforts well short of what is needed. As a result, it is permanently out of touch with its customers' needs. Government should confine its extension activity to provision of information through broadbased multimedia systems, while customer specific activities should be privatized. Otherwise, it is difficult to see how to avoid regular recurrence of such identity crises for the extension service. Placed on a firm commercial footing, all parts delivering value added to commercial agriculture would flourish, while those not needed would be retrenched. In research, wholesale privatization may not be desirable. But, interaction with commercial, demand-oriented extension companies would greatly aid in improving the focus and impact of research. 21. Private Extension and the Subsistence Farmer. At first glance, privatization of extension services for subsistence agriculture may not appear feasible. Although it could help them, and so be socially desirable, individual farmers may be unable to see the benefits and so unwilling to pay for them. Even where the benefits are seen, cash-flow problems may limit access to this input. Government could very easily overcome this via a contract with private extension companies to deliver these services to subsistence farmers, initially at a subsidized rate, as part of an overall development package for subsistence agriculture. 22. The Growth Challenge - Getting Agriculture Moving! Simply put, this is the fundamental challenge facing Sri Lankan policy makers. Years of policy induced stagnation have made the growth challenge facing the sector and thus the economy very great. For sectoral growth to move from 2% to a modest 3.5% p.a., output from paddy land would need to grow year in and year out by about 5 % p.a., i.e. over 20 times its present rate (assuming the growth of OFCs remain the same as now and plantation crops grow at 1 % p.a.). This is simply not possible for paddy itself. Instead, to get agriculture moving the cropping pattern will need to change in response to a corrected incentive structure. 23. Government - a new role in commercial agriculture? What should be Govermment's contribution to the sector and in what form should it be made? To achieve the needed growth in a sustained manner, Government should rethink its role in the sector. Its traditional "interventionist" approach no longer has the impact needed. It is quite clear that direct market interventions, regulations or appeals are no longer of any value if these are not in the best personal interests of farmers in general. Instead of calling for the desired results, Government will need to use the signals to which farmers now respond and set these so as to get the needed results. Government will need to adopt a more distanced stance relative to its current "hands-on" mode and to operate more in the background, leaving the actual production and marketing entirely in private hands, including the allocation of inputs such as credit and land. 24. Government - a new approach to subsistence agriculture? Alongside this recornmercialization, direct Government programs to address the needs of subsistence farmers must also be priority. At the core of this are carefully targeted programs designed to relieve constraints to income generation for those who choose to remain in agriculture, and assist those who wish to exit agriculture to do so on the best possible terms. For individual persons exiting agriculture, it will be through expansion of off-farm employment by ensuring adequate infrastructure investment. For those remaining in agriculture, this will be by increasing productivity from enlarged farm size (from land freed as others leave agriculture) and improved farming abilities. 25. Priorities for Agriculture: Priority actions to get agriculture moving again are: initiation of trade policy liberalization, adopting a policy of full private ownership of agricultural land and fully operational land markets, implementation of land privatization and land market development programs, commercialization of the irrigation system, through creating water property rights and markets and a system of management by owners, phasing out of all Government sponsored market intervention programs (including privatization of the PMB, CWE, all public sector financial intermediaries, and the commercial activities of the Food Commissioners Department). This will allow mainstream agriculture to recommercialize. Generally, agricultural sector administrators should take the approach of their colleagues in industry and trade. 26. Sequencing Reforms: At present lower import protection and increased irrigation costs, which would reduce paddy profitability, may also reduce output. Without a well functioning land market, land utilization would decline while strong off-farm labor demand would absorb the surplus farm labor. Rural income may increase, but agricultural output and income would decline. But access to long-term credit is a prerequisite for smallholders to access land markets and so expand. Long-term credit requires marketable collateral (titled, freehold land) and enforceable credit discipline, which calls for institutional development in land titling and recording, as well as change from current ad hoc loan forgiveness interventions in agricultural credit to assistance in enforcement of legally binding loan contracts. Uncertainty in irrigation policy on water property rights and charges introduces uncertainty in land valuation, which willV depress land markets. Finally, introduction of some taxation may be necessary to equalize any differences in asset acquisition costs between the nonplantation sector and other sectors. Otherwise, any difference will be capitalized into V the market value of land, again making it difficult for smallholders to expand farm size. Any program to address the problem must take these aspects into account in designing the sequence of reform. 27. Governiment's Preparedness: Government is well aware of the challenge it faces in revitalizing the nonplantation sector. In this regard, a major policy speech' noted the requirement for "substantive structural change", while noting Government's "confidence in the rural masses and their ability to make rational choices". The degree of alignment between this statement of policy intent and the findings of this study shows a sound appreciation of the nature of the situation by Government. With the exception of agricultural markets, where Government intervention rather than structure or conduct of the market is the problem, the study's findings are in complete agreement with Government's policy intent. It is hoped that the study will provide greater depth to understanding the current situation and thus a basis for improved design of reform policies and measures. 1/ Policy Statenient ot the Government of Sri Lanka on the Occasion of the opening of Parliament, 6 January, 1995, by H.E. Chandrika Bandaranaike Kuniaratunga. President. SRI LANKA NONPLANTATION CROP SECTOR POLICY ALTERNATIVES I. THE PROBLEM 1.1 Poor overall performance, shown by the very low growth rate in sectoral income, is the central problem. In the plantation sector the reasons for this are now quite well understood'. But, up to now, the determinants of performance in the nonplantation crop sector remain unexplained. 1.2 Agriculture is still one of the largest sectors in the economy, with 18% of GDP (excluding forestry and fishing - 3.5% of GDP), about 45% of employment and 24% of gross exports in 1992. Sectoral growth has been poor, about 2% p.a (constant terms) over the past 10 years (1982-92), or half the national average. Some 65% of the population live in rural areas and are seen as largely dependent on agriculture for their livelihood. 1.3 Allocation of land for nonplantation crops is mainly for Paddy, a low valued commodity. Although average income per acre from Other Food Crops is three times that of Paddy, and, since 1980, the international price of rice has declined, in real terms, by over 6% annually, this allocation has remained constant. Domestically, the average annual decline of 1.4% in the domestic producer price over the past 10 years, has barely been offset by yield gains. This may explain why the sector, with 25% of total employment, still only receives 12 % of GDP. But the question of why or how such resource allocation emerged, and what sustains it, unanswered thus far, is central to explaining sector performance. Land tenure policy and conditions in the off-farm labor market emerges as pivotal here. Sectoral Goals aiid Strategies 1.4 The National Agriculture, Food and Nutrition Strategy (NAFNS)2, published in 1984 remains the sole comprehensive sectoral policy document available. It took as a starting point GOSL's four sectoral objectives, namely: * Self-sufriciency in basic foods, rice, milk, sugar, fish and pulses; _ Expansion of export capacity to increase agriculture's contribution to the balance of payments; * Enhancement of incomes and creation of new employment opportunities in the rural sector; and - Improved nutritional status of the people. 1.5 Implementation strategies in the NAFNS document for the nonplantation crop sector centered on paddy. With rice self-sufficiency perceived as virtually attained, the concern was disposal of surpluses. Substitution for imported wheat, 'iversification of land out of paddy, and concessional supplemental nutrition programs were all proposed. Substantive progress towards sugar self-sufficiency was seen as a major near-term objective. Agroindustry was seen as "a key element in generation of off-farm employment and the basis for rural industries'. A conducive policy atmosphere was seen as essential here. In development of "Other Field Crops' (almost all 'Other Food Crops") 'the traditional in-built bias in favor of rice' was noted as a major inhibiting factor to progress in diversification. 1.6 As with many small island countries and isolated areas, the history of Sri Lanka is bound up with the continuous struggle for survival, particularly the struggle to ensure adequate food supplies. Consequently, it is normal to find that all facets of rice, the staple food of most Sri Lankans, are a major focal point of the socio-economic history. For centuries, "temple and tank" (irrigation) were the focal point of rural life. History illustrates how a ruler's worth was measured in no small part by his contribution to these two elements3. This perspective appears unchanged even today. Avoidance of dependency on volatile international markets drives the pursuit for food self-sufficiency. Investment in large-scale irrigation schemes has both supported this pursuit, and created land for the landless. I/ See Sri Lanka Tree Crops Strategy, World Bank Report No. 12356-CE, July 5. 1994. 2/ National Agricultural and Food Strategy A Change in Perspective. National Planning Division, Ministy of Finance and Planning, Colombo, Sri Lanka. June 1984. 3/ See "Food and [he People" R.L. Brohier, Lake House Publishers, Colombo, 1975. 2 1.7 Overall, the strategy document felt that "There are good prospects to raise incomes and improve nutrition of smiallholders through removal of constraints to productivity, wider opportunities to produce and market their crops, and a more versatile and adaptive institutional framework to execute the required changes." The NAFNS document was not official Government policy, rather it was developed by a central Government agency and "presented for further discussion and review". L .8 Following World War 11, the basic elements of current sectoral intervention by Government became a permanent feature and progressively institutionalized. Direct Government intervention has been the strategy of all governments over the past 30 years in pursuit of sectoral policy objectives; rice self-sufficiency being the primary one. In this scheme, domestic producer and consuiaer price policies are the central focus of medium-run policy measures, while in the longer run fixed assets, particularly land, have been the focus of intervention. Trade policy was designed to be supportive of domestic price policy. 1.9 Policy discussion in response to declining sectoral performance, has ranged between promoting commercial, export-oriented agriculture and continued protection of a peasant-structured agriculture. Depending on the ministry (central or line) the call has been for more of the former or the latter; but sustained action has been negligible. On one hand caught between the populist sentiment embodied in the policy objectives and the high economic cost of the interventionist strategy required to pursue them, and on the other between the desire for improved sector performance and the possible political cost of pursuing this, Government has been unable to make any direct headway in rationalizing sector policy. With past policy measures now a tradition, the social dimension has overridden any commercial dimension to Sri Lanka's agricultural policy. There may be little conviction in the appropriateness of the current objectives and strategies, but there appears an equal inability to recast them. Unable to sustain high cost irrigation investment and with land reform possibilities exhausted, the objectives pursued implicitly narrowed to a more affordable, yet still politically acceptable set, namely: o Food Security - redefined as "rice self-sufficiency at an affordable cost"; * Cointrolling cost of living for the poor; and - Maintaining farm income. 1. 10 Pursuit of these objectives involved continuation of Government's dominance in the economic life of the sector, through direct regulation of price levels and land use, state trading, and direct control over trade. While this was a continuation of earlier policies, over time there was an easing of legislation directing land use, and a reduction in state trading with a concomitant relaxation in enforcement of Government set producer prices. Direct intervention in trade, although changed in character due to the reduction in state trading, remained as active as ever, with no change in the disruptive nature of its ad hoc character. The primary motivation for these changes appears to have been the need to control Government spending rather than improve sector performance. Reduction in state trading meant reduction for Treasury in the trading losses and financial charges it faced on behalf of the state trading agencies. With indirect intervention replacing the direct approach, underlying policy remained unchanged. 1.11 Despite Sri Lanka's good social indicators, this affordable-cost approach does not appear to be working. The opportunity cost of defining "food security" in terms of "physical self-sufficiency" may actually be unaffordably high in terms of sectoral growth foregone. Certainly the current level of "food security" was purchased at a high cost through investment in very capital-intensive irrigation schemes to produce a low valued commodity. This lack of growth in the rural sector also constrains both direct and secondary growth in income and employment. 1.12 The new Government, elected in August 1994, in fulfillment of an election commitment, reduced some food prices, particularly bread and rice. This was accomplished through increased import of wheat and resale as flour at subsidized prices by the CWE (Cooperative Wholesale Enterprise), a major state trading agency. The negative impact on the domestic rice price required increased purchases by the PMB (Paddy Marketing Board), which was also required to sell rice purchased at below cost as part of the food price reduction effort. In addition, since the wheat flour subsidy was untargetted, its effects were regressive with the 20% highest income household receiving 32% of the subsidy and the 20% lowest income households receiving only 9%. Subsequently, Government indicated that these measures were temporary and that "a natural balance will be restored to the market place". 3 The Policy Dilenmma 1.13 Despite its low returns per acre relative to other food crops and Government support for more diversified crop- land use, low-valued paddy remains the dominant crop. Why this is so has been a dilemma. Although net revenue per acre averages about four times that of paddy', with no growth in the 12% of cropped land allocated to Other Food Crops (OFCs), diversification has been highly concentrated on a small area. With very little distributive element, the impact on poverty alleviation is limited. 1.14 Diversification has been a policy of Government for sometime now, as manifest in the Amendment to the Agrarian Services Act discontinuing the restricted use of paddy land, and the gradual development of a research and extension capability in this area. Despite this, and the superior farm income generated, actual diversification of land use is negligible, as growth in output of Other Food Crops has been intensively rather than extensively based. This output stagnation is a major concern to Government policy makers, and its cause, though frequently debated, remains unidentified. 1.15 There appears to be some capacity for further expansion in output from traditional 'Other Food Crops' (i.e. import competing crops, primarily chilies, onions and potatoes), in view of the protection afforded through the trade regime. Although the present economic viability of these is questionable, creating an additional dilemma for policy makers, this does not explain the static allocation of land to production of these. Seasonal price fluctuations here are likely to be relatively large. This would incline risk averse farmers towards lower, but more stable, income from paddy production. Nonetheless, the data indicates that existing producers not only continue in production but are actually intensifying output effort on the same area and thus risk. Also, the continuing decline in income from paddy could be expected to at least partially offset aversion to risk. Finally, the impact of the somewhat arbitrary import regime for OFCs needs scrutiny to see if this source of price fluctuation can be reduced5. 1.16 The study systematically examines the performance of the concerned product and factor markets to determine their contribution to current sectoral performance and identify elements which might impede adjustment and growth. The report summarizes the detailed product and factor market assessments and interprets these in the light of the policy makers dilemma concerning policies needed to improve sectoral performance. Specific markets examined are paddy/rice, and major OFCs (chilies, potatoes and onions) on the output side, and on the input side, purchased inputs (seed and fertilizer), rural finance, and land. U. MARKETS' OutDut Markets Rice 2.1 Structure. With over 800,000 mostly small farmers selling to numerous independent traders and millers (over 7,000 in 1986, according to a PMB survey), who in turn supply milled rice to consumers, the domestic market for paddy/rice is basically competitive. Because of the relatively low level of technology used, there is little market concentration at this stage, and charges for storage and distribution functions are reasonable, indicating competitive pricing. Information on prices is readily available, and there are no significant barriers to entry. Inputs for paddy/rice trading such as trucks and casual labor are also readily available. The credit market is ready to provide both working capital and investment financing for mills and stores. Bank financed industrial credit lines have provided substantial input financing for these. 2.2 Conduct. Although structurally competitive, the operation of the paddy market is heavily influenced by Government policy. Government's main intervention arm is PMB. It is responsible for administering the Guaranteed 4/ Cost of Cultivation of Agricultural Crops, Maha 1992/93. Socio-Economic & Planning Centre, Dept. of Agriculture, Peradeniya, March 1994, Tables 8 & 18. This factor varies from 3 to 7 depending upon the crop, year, and data base. 5/ Imports by state tnading agencies in high production season, a not infrequent occurrence. would not occur if the importer was a stricdly commercial entity. 6/ Detailed analysis of performanlce of the domestic markets is in WP#I. 4 Price Scheme (GPS), even though, for reasons of its small market share (25,000 tons annually or about 2% of the marketed quantity of paddy over the period 1989-93), it has been powerless to guarantee a price to producers. The fact that producer prices have been below the "guaranteed" price in the main harvest months confirms that PMB is no longer effective in carryiig out its main function. Also, given that PMB purchases paddy delivered to its stores and that most of its purchases are from private traders and cooperatives, with only a small amount coming directly from farmers, it is likely that most of the margin between the "PMB" price and the "free market" price is captured by traders rather than going to producers. But activity in the wholesale market competes away much of this margin to the retail market and so to the consumer. 2.3 The other major Government lever over the rice market is tariff policy. Present policy is for imports to be made by the private sector, either by normal licensed traders or by licensed 'bondsmen", who are pennitted to import without paying duty until their rice is released onto the domestic market from bonded warehouses. Since November 1993, an import duty of 35 % (on a standardized import price of US$225/ton) has been maintained. In early December 1994, the rate was reduced to 20%. 2.4 Performance. PMB's ineffectiveness in influencing prices has not been without substantial costs. Over time, PMB has incurred a heavy penalty for the relatively high prices which it has paid for paddy. For instance, in 1994, the wholesale market price for rice tended to be consistent with the open market price of paddy and in consequence PMBs own high procurement price plus costs associated with handling, storing and milling means that it has not been able to cover its costs for rice when selling on the local market. Under instructions from the Government, in the second half of 1994, PMB sold at least part of its stocks of rice at a price of Rs. I 1/kg (US$220/ton), resulting in a loss of about Rs.4/kg (US$80/ton). PMB was expected to lose Rs.2/kg (US$40/ton) on its total rice sales during 1994. Total operating losses for 1994 were seen to be in the order of Rs. 150-200 million (US$3-4 million). 2.5 The wholesale price of rice in late 1994 was too low for imported rice to be released profitably onto the domestic market, although because of the need to rotate stocks, small quantities of imported rice have been sold by bondsmen at a loss, both to the Cooperative Wholesale Establishment (CWE) and through the Pettah market wholesalers. In December 1994, imported rice was being offered at a wholesale price of Rs. 16.75/kg (US$335/ton), but was meeting slow demand. To have fully covered the bondsmens' costs, a price of at least Rs. 19/kg (US$380/ton) or Rs. 17.3/kg (US$346/ton) with the new lower tariff, would have been required. This situation, which is not commercially sustainable, results from: (a) a relatively high level of domestic paddy production; (b) a currently buoyant international market for rice; and (c) the availability of heavily subsidized flour and bread which are suppressing demand for rice. 2.6 Policy Conclusions. The Government's rice market policy has failed to achieve its objective of supporting producer prices, and has prove very costly to the Government. It has also resulted in higher-than-world prices for rice to consumers. 2.7 Prowosed Actions. Local production should be protected by a moderate tariff (10-20 percent). All quantitative import restrictions should be dropped. Guaranteed minimum producer prices should be removed, PMB and CWE privatized, and the commercial activities of FCD discontinued. Wheat .2.8 Structure. Unlike rice, the wheat market is basically non-competitive, i.e. a Government monopoly. The amount of wheat imports, accounting for 100% of the country's flour requirements in 1994 and 1995, is administratively determined to achieve a fixed domestic price. Even though the private sector has been allowed to import 40% of the total (since the beginning of 1994), almost all imports are through the Government-owned CWE. Milling is handled exclusively by a single private firm, PRIMA, which has a 25-year contract with the Government which began in 1980. 2.9 Conduct. Under the PRIMA contract, the Government is obligated to provide 435,000 tons of wheat annually for milling. For each ton of wheat milled, PRIMA must return about three-quarters (740 kg) in the form of flour (or alternative products such as semolina) in exchange for being allowed to keep the remaining wheat bran and other by- products of milling for export as animal feed ingredients. In practice, PRIMA has almost twice this capacity and mills the totality of the country's imported wheat, producing only one type of all-purpose flour made from 50% hard and 50% soft wheat. Virtually all of the output from the PRIMA mill is provided to CWE which has the flour distributed by the Food Commissioner's Department (FCD). The mill is fully amortized, and at the expiration of the contract in 2005, it is expected to be handed over in good working order to the Government free of charge. 2.10 Performance. Until August 1994, the fixed price of flour had been slightly above CWE's costs, and over the years CWE had been able to make a modest profit from flour trading. This has now changed with the price of flour 5 being reduced to Rs.7/kg or about 60% of its current cost of production. The loss incurred is currently being absorbed by CWE, with the official explanation that CWE is sharing past profits with consumers. In reality, CWE's past food trading profits (about US$12 million) have already (December 1994) been absorbed by the flour loss. Future loses will need to be very quickly recompensed if CWE is to remain a viable and credit-worthy organization. Estimates are that 1995 losses from continuing with this program are likely to be of the order of Rs.4 billion (US$80 million), more than five times CWE's equity at the end of 1993. 2.11 Flour is distributed at a standard fixed price throughout the country by the Food Commissioner's Department (FCD) which has over 50 warehouses with a total storage capacity of 500,000 mt. Only 40% of this capacity, however, is utilized. Some flour in FCD's stores is also reported to be in poor condition (clumpy and variable in moisture content), reflecting difficulties in inventory management. This is disposed of as animal feed. Approximately 90% of the flour is distributed to some 7-8,000 Multipurpose Cooperative Societies (MCPS), while 10% goes to 8,000 bakeries. The price of the 450 gram loaf of bread is officially unregulated, but bakers allege pressure to maintain the price at a level which does not cover their costs, resulting in a poor quality product. The prices of other flour products such as noodles, pasta, pastry and other types of bread are not controlled. The attempts to control the price of flour and the 450 gram loaf are intended to hold down the cost of living, but are not part of the targeted food subsidy system which relies on food stamps. Flour and flour products account for about 9% of the family food budget, second only to rice at 23 %. 2.12 Perhaps the biggest concern over the wheat and flour policy is that, although Sri Lanka is a country which grows rice and imports wheat, the Government is adopting a contradictory policy whereby consumer demand is being artificially suppressed for rice by having a policy of.protection, but encouraged for wheat and flour by direct subsidy. The budgetary cost of this operation is substantial. Because Sri Lanka is now near to rice self-sufficiency (and indeed in surplus for much of the year), this combination of policies is likely to lead to increasing volatility in domestic paddy prices. When exports are needed to clear the market, prices would fall to the adjusted world price less the transport costs, yet at times of shortage the domestic price would rise to the world price plus transport, plus duty. 2.13 Policy Conclusions. The present heavily public sector controlled system of wheat marketing has led to a variable quality product with a high cost to the budget. Although it has benefited consumers through lower prices for wheat products, it has created distortions in the substitute rice market which contradict the Government's rice price policy. 2.14 Prowosed Actions. Commercial imports of wheat and flour should be opened to all interested parties in the private sector, while imports of wheat and flour under Government-to-Government arrangements (e.g. US PL480) should be auctioned. It is understood that these arrangements would apply after the Government has satisfied its contractual commitment to PRIMA. The domestic price of flour should no longer be subject to any form of control, and distribution at the wholesale and retail level should be fully liberalized. The facilities of CWE and FCD concerned with wheat marketing should be privatized. To facilitate private marketing of wheat, paddy millers should be encouraged to investigate the possibility of investing in small-scale flour milling equipment. Other Food Crops (OFC) 2.15 Structure. Maize, pulses, chilies, onions and potatoes are largely marketed on a competitive basis by the private sector, but with the partial exception of chilies, there is no intermediate processing stage. The raw products are procured in the rural areas by assemblers who package them in gunny bags and either sell them via the Pettah (Colombo wholesale market) commission agents, or deliver them to "polas" (local town and village fairs) where they are sold to consumers via private retailers. Because some of these products are produced by larger, more commercial farmers, particularly potatoes and onions, significant quantities move directly from farmer to wholesaler, bypassing the assembler. Storage is undertaken by farmers, assemblers and wholesalers. Commissions in the Pettah market for these commodities are around 5 %, compared to 3% for rice, and retail margins are also greater than for rice. 2.16 Conduct. Direct Government intervention in the OFC markets has been minor. However, because there is officially a guaranteed price scheme covering maize, some of the pulses and grains, and until 1992, dry chilies, some purchases have been made by PMB. The main intervention was in 1979 when PMB purchased about 10,000 tons of maize or about one-third of the total crop. In most years, the market price has been significantly above the FPS set price. CWE has also purchased some of these crops on a commercial basis. The most significant involvement was in 1992 when CWE purchased some 15% of the big onion crop and 12% of green gram supplies. It also became involved in big onion trading in 1994, although this was a specifically Government-inspired intervention, and so was not entirely on a commercial basis. Overall, only about 3 % of OFC production has been handled by PMB and CWE. 6 2.17 The main influence by Government on OFC markets has been through trade, via both policy measures and direct trading. Imports of chilies and onions are subject to quotas which are so restrictive that they ban all imports in some years. The quotas are imposed in an ad hoc manner, generally during the domestic harvesting season. For potatoes, the policy does not allow any imports, although imports of seed potatoes are allowed under a licensing scheme. 2.18 In addition to quantitative restrictions, imports of chilies, onions and potatoes are subject to tariffs. For onions, the ad valorem rate is 35 percent and the specific duty Rs.9/kg. The highest tariff applies. For chilies, the ad valorem and specific rates are 35 percent and Rs.20/kg, respectively. For potatoes, they are 35 percent and Rs.12/kg. Imports of chilies and onions have been a monopoly of CWE; but since 1993, private traders are allowed to import up to 40% of the total. 2.19 Performance. Government's policy of quantitative restrictions, high tariffs and direct trading in onions and chilies during certain periods of the year, together with the effective exclusion of the Jaffna region's production from the Colombo market, has meant that prices have been quite volatile. Very low prices at harvest time have been followed by rapid price increases as domestic supplies run down, but imports are banned. For example, the wholesale price of big onions, reported by ARTI to have averaged Rs. 12/kg (US$240/ton) in September 1994, had increased to Rs.261kg (US$520/ton) by the fourth week of November and to over Rs.30/kg by early December (US$600/ton). This price rise is substantially greater than the likely cost of storage, indicating that there are some discontinuities in the market. 2.20 Part of the reason for the high price volatility associated with these crops is that market participants find it difficult to be confident about the environment within which they trade. Traders fear that Government may rapidly change the rules and naturally, they try to position themselves so that such changes do not impose a heavy financial burden. One way of doing this is to hold minimal stocks, thereby making storage unattractive. A case in point occurred in late 1994 when onion, potato and chili farmers and traders were led to believe that no imports would be allowed until January 15, 1995. However, the Government subsequently changed its mind and liberalized imports as of December 7, 1994. While economically sound, this sudden reversal unexpectedly reduced the profitability of investments in storage, causing losses for some and reducing future willingness and ability of private traders to invest in storage. 2.21 Policy Conclusions. Direct trading by CWE and PMB in OFC combined with quantitative restrictions on OFC imports constitute unnecessary interference in a basically competitive market where traders would normally even out price fluctuations through storage, given a more neutral policy environment. At different times, producers, consumers and would-be investors in storage have been the victims of the resulting price volatility. 2.22 Proposed Actions. Importation of OFC should be liberalized, and made subject to a moderate tariff (10-20%) only, with no nontariff restrictions. CWE's and PMB's activities in the OFC markets should be discontinued. Fruits and Vegetables 2.23 Structure. Marketing of perishable fruits (mango, lime, papaya, passion fruit, plantain bananas, pineapple and orange) and vegetables (cabbage, tomato, beans, carrots, radish and leeks) is handled entirely by the private sector. The main clearing market is the wholesale market at Pettah in the heart of Colombo. Traffic here is very congested, making access difficult and so increases transaction costs. Wholesale stalls are leased from the municipality at a low rate, but there are substantial barriers to entry with rents being garnered by the long-established stallholders who only pass on their businesses within the family. The market is thus competitive, but within a restricted context. Alternative informal sites are developing outside both Pettah and Colombo itself; but these are still quite small. 2.24 Conduct. Mark-ups between the Colombo wholesale price and retail price appear very high for vegetables, but reasonable for fruit. One of the reasons for the apparently large mark-up for vegetables is the very high level of physical losses incurred at the retail level. Vegetables are brought from their production areas into Colombo in sacks which are roughly handled and piled high on non-refrigerated trucks. They are wholesaled within the same containers in the crowded Pettah market; consequently, it is only when the retailer unpacks his purchases that the transit losses are evident and reflected in the unit price. With fruit, which is largely retailed by the piece, much of this type of loss is reflected differently in lower unit weights or quality to the consumer. 2.25 Performance. A review of margins for individual commodities suggests that with low value added products, retailers attempt to make a minimum spread per kg sold, while minimizing price fluctuations. So, for example, at the end of September 1994, when cabbage was very cheap and wholesaled at Rs.4/kg, the retail price was Rs. 18/kg, giving- a spread of Rs. 14/kg and a percentage mark-up of 350%. By the beginning of December, the wholesale price of cabbage had risen to Rs. 13/kg, but the retailers add on had increased only to Rs. 15/kg, giving a percentage mark-up of 129%. 7 2.26 Policy Conclulsions. Even though margins within the retail vegetable trade do not appear excessive, given the perishability factor, one way in which these could be reduced would be to enlarge and improve wholesale facilities to reduce the large physical losses presently being incurred. Moving the wholesale market from central Colombo to a new site with larger purpose-designed buildings and offering the new stalls and other facilities to be bid for by competing wholesalers would also help reduce barriers to entry and stimulate competition. Both producers and consumers would benefit. 2.27 Proposed Actions. Develop project proposal for donor financing of a new market in a new location. Input Markets Fertilizer 2.28 Structure. With 15 registered importers, a range of distributors and distribution points throughout the country, including the Agrarian Service Centers (ASCs), and adequate and relatively inexpensive trucking, the fertilizer market is structurally competitive. Government owned industries used to dominate the industry, but the situation has changed rapidly, partly as a result of the sell off of the Government firms' assets. Over 60% of imports are now in private hands. 2.29 Conduct. Although prices are nominally fixed, the availability of significant discounts for quantity, which are decided on a company by company basis, means that in reality there is price competition. Prices charged by one of the companies a]so differentiate between location of supply by amounts which are generally consistent with transport costs. Clients are also given the option of lowering their costs by using recycled sacks. 2.30 Performance. Supply problems have arisen, but these are due to Government intervention rather than to market structure. For example, in an attempt to maintain fertilizer prices at a constant level to the farmer, the Govemment restored the subsidy which was removed in 1990, but all indications were for further changes, including a possible voucher system7. If importers or traders believe that fertilizer prices will go up in the near future, it would be rational for them to withhold supplies at the moment with the expectation of selling them at much higher prices in the future. A further cause of shortages is that Japanese aid fertilizer, which is normally delivered in time for the Maha season, and is distributed at a more subsidized price than other fertilizer, has been delayed. 2.31 Policy Conclusions. Evidence shows that fertilizer consumption regained and even surpassed the levels achieved when fertilizer was subsidized in the late 1980s, so that it is not necessary to subsidize fertilizer to ensure its use. Sudden shifts in subsidy administration are also disruptive for both producers and consumers and may lead to rent- seeking behavior. There may also be environmental reasons not to stimulate fertilizer consumption. 2.32 Proposed Actions. The fertilizer subsidy should be eliminated. Grant funded fertilizer should be auctioned to the public. Seeds 2.33 Structure. In 1991, the seed market consisted of 46% paddy seed estimated at Rs. 1.1 billion (US$22 million), 38% potato seed worth about Rs.0.9 billion (US$18 million) and 17% other field crops and vegetable seed valued at around Rs.0.4 billion (US$8.0 million). These figures probably underestimate the 1994/95 value by some 30%. They also exclude other field crops which might be important for an emerging private sector seed industry such as cotton, sunflower, maize and soybean. Seeds used may be either the farmers' own retained seed (47% by value), seeds purchased from another farmer (21 %), seeds procured through formal suppliers (31 %) or seeds from 'other' sources, e.g. landlords, or merchants to whom crops are contracted (I %). 2.34 The formal supply system for seeds in Sri Lanka is partly private and partly public. Historically, paddy seed has been handled by the public sector, the production of which (both on their own farms and under contract) peaked at about 15,000 tons in the mid-1980s, but now amounts to only about 4,000 tons or less than 5% of total gross needs. 7/ This subsidy was replaced in April 1995 with a scheme which puts a ceiling on the budgetary cost and limits its application to four hasic fertilizers, Urea, ainmmonium sulphate, muriate of potash, and triple super phosphate. Based on announced subsidized prices per 5t)kg sack, the effective rate of subsidy is now: urea 40%, ammonium sulphate 15%, muriate of potash 9% and triple super phosphate 22%. The subsidy is patid to importers. Increases in import prices will be passed on to farners to maintain the budgetary ceiling. Decreases in world prices will be deducted from subsidy payments, reducing the budgetary cost. 8 The level of certified seed is believed to be far below the optimum level. Although paddy is not a hybrid and so seed can be reproduced at the farm level without specialized technology, the need to maintain purity and regularly introduce new improved varieties suggests that about 20% of all paddy seed planted each year should be certified seed. Private business has been allowed to participate in seed paddy production since 1990, particularly in the production of certified seed from registered seed. However, the rate of growth has been limited and private production of certified seed is unlikely to amount than more than 1 % of total seed in 1994/95. 2.35 Conduct. One of the reasons for the slow development of private sector supplied seed is Government's seed registration policy. Government requires that varieties be approved and registered before they can be sold. Another reason is what is seen by the private sector as Government selling seed well below the cost of production thereby constituting unfair competition. A recent announcement by the Department of Agriculture that they are to reduce their certified paddy seed prices by a further Rs.50 per bushel (about US$50/ton) for the 1995 Maha season planting has caused at least one large private sector paddy seed producer to plan for a drastic cut-back in his own production and marketing plan. The Ministry of Agriculture, Lands and Forestry (MALF) is not in agreement with this conclusion, believing instead that private sector seed companies are insufficiently committed to providing paddy seed in a manner which meets farmers needs.8 2.36 For high-valued vegetable seeds, the market is largely in the hands of the private sector, which distributes through agents and sub-agents in the districts. Many vegetable seeds are imported, as too are seed potatoes. The latter situation results mainly from the anomaly whereby importation of potatoes for consumption has generally been banned whereas importation of seed potatoes is permitted. This has created an incentive to import consumption potatoes under the guise of seed potatoes. 2.37 Performance. Despite this peculiarity, the predominantly private sector operated system of seed production for high value vegetables appears to work reasonably well. Analysis of Government's performance in paddy seed production, however, has indicated operational inefficiencies. 2.38 Policy Conclusions. The general policy of Government in this area should be to intervene only in areas of market failure. Government financed plant breeding should be restricted to nonhybrid seeds only and all testing, registration and certification should be aimed at overcoming any problems of asymptotic information in the domestic seeds market. Where Government believes that commercial seed is not as advertised and farmers are generally unaware of this, it should carry out tests on the seed and make the results widely known in the seeds market. In the interests of developing a private sector seed industry capable of supplying high quality seeds to farmers at reasonable prices, Government's involvement in paddy seed should be confined to the research linked areas: adaptive breeding; carrying out variety trials; testing, registration and certification, but only as a voluntary service to farmers and seed dealers, who should be free to buy or not to buy Government registered/certified seed. The private sector should undertake all commercial seed sales and distribution, so as to avoid any possibility of unfair competition from public sector seed companies whose managements are not faced with hard budget constraints and financial self sustainability. There should be no restrictions on import of seeds other than for minimum basic phytosanitary reasons. Otherwise, Government policy is likely to be the root cause of a major constraint in access to new technology by the sector. MALF indicates a general acceptance of this policy approach and claims to be following it. III. INCENTIVE POLICY Price, Trade and Marketine Policy 3.1 With its central role in food consumption, intervention measures center on the farmgate price of paddy. A Guaranteed Price Scheme (GPS) was introduced for paddy in 1948, which set the domestic producer price above the international equivalent price. Implementation was through direct Government intervention via a Government purchase program (the Department of Agrarian Services until 1971, and thereafter the Paddy Marketing Board (PMB). The GPS is still in force, though in recent years, events have made it insignificant. In the late 70s, the GPS was extended, in the form of a Floor Price System (FPS) to cover some OFCs, and dampen domestic price fluctuations. But, due to policy induced decline in the activities of the Cooperative Wholesale Establishment (CWE) and PMB, its impact appears has been superseded by the underlying restrictive import policy. 81 Tue difference of opinion here arises from differing perceptions/definitions of farmers needs and leads to the critical question of the financial sustainability of MALF's tiwn paddy seed supply service. 9 3.2 Up to 1978, marketing of rice and import of rice, potatoes, chilies and onions were very much under direct Government control. At its peak, Government, through CWE and PMB, purchased 64% of output in 1965, which, together with imports by CWE, were distributed to consumers by the Food Commissioner's Department (FCD) through Multipurpose Cooperative Societies (MPCs) and authorized private agents, enabled Government to control (keep down) open market prices. Rice rationing, first introduced in World War 2, was reintroduced in 1960 to redistribute the domestic production as a result of difficulty in access to foreign exchange and so imports. For the following 7 years, as much as 80% of all rice consumed annually was rationed rice. Subsequently, this declined to about 50%, as a new Government reduced the rationed allowance in response to the budgetary strain (17% of Government expenditure in 1975) of a rising program cost, due to rapidly increasing world food prices; as compensation, Government did eliminate all charges for the reduced ration. 3.3 Following a change in Government in 1977 came substantial liberalization of the economy in favor of outward orientation and reduced Government activity. Aimed mainly at industry, it left agricultural price and trade policy fundamentally untouched. But there was a marked reduction in the aggressiveness with which the policy aims of the previous Government were pursued. Govemment allowed the real level of protection from competition by imports to erode and Nominal Protection Coefficients (NPCs) for rice9 fell from 1.5 in 1976 to 0.7 in 1979 (0.49 when exchange rate adjusted). Also, Government intervention in paddy marketing was greatly reduced, so that purchases by PMB fell from 32% of output in 1978 to 4% by 1981 and remained at or about this level thereafter.'" Incentive Coefficients - NPCs. EPCs and ESCs 3.4 Overview. As a rule, sectoral policy incentives are set by Government intervention in the prices of traded goods, i.e. outputs and purchased inputs, via trade measures (tariffs, quotas, licenses, etc) and direct subsidies, e.g. on outputs and purchased inputs such as fertilizers. On the output side, these trade protection effects are measured by the NPC or nominal protection coefficient, while for the output and input sides combined, the amount of protection is measured by the EPC or effective protection coefficient. But in Sri Lanka, the big story is in the subsidies given to nontraded inputs, particularly irrigation water. The effect here is measured by the ESC or effective subsidy coefficient. This is the deepest measure of protection. It is generally not a major factor in explaining sectoral performance. But where Government has financed major irrigation works, and in the absence of any cost-recovery, this can become a major factor, as it has in Sri Lanka. 3.5 NPCs. In recent years (1985-93) NPCs" for the main import-competing crops (paddy, chili, onion and potato) were well above 1. The NPC for rice, the main non-plantation crop in Sri Lanka, averaged 1.2 for the entire 1985-93 period, though showing an upward trend in 1990's. It is now about 1.3. This indicates that on average the barriers to rice imports held the domestic price at 20% above the import price. The NPC for wheat, the main substitute for rice in consumption,'2 averaged 1.0 for the entire period, except for late 1994, when consumption of wheat flour was granted a subsidy of about 40%. For 1985-93 period, NPCs averaged 1.3 for chili, 2.0 for big onions, and 1.6 for potato. On average, onion is the most protected crop, followed by potato and chili and then rice. Between the two subperiods (1985-89 and 1990-93), average protection for rice and big onion has risen, whereas it remained unchanged for chili and potato. Using production weights, the average NPC for the import competing crops works out to be 1.33 in 1993, which is very close to NPC for rice (1.3). 3.6 EPCs - Import Competing Crops. Effective Protection Coefficients (EPCs) measure the net incentive provided to the production of a crop by taking account of the protection or disprotection received by the traded inputs used in its production. The protection provided to traded inputs (fertilizer, agrochemicals, farm machinery etc) since 1990 was small. The Maha 1993 EPCs for paddy (estimated for different districts with different water regimes) range between 91 Derined an the ratio of domestic price over border equivalent price. 1O/ Economic Policies and Agricultural Performance in Sri Lanka 1960-1984. Eric Thorbecke and Jans Svejnar, OECD Paris, 1987, Tables 28 & 29. II/ For details on estimation of incentive and comparative advantage coefficients and review of other estimates see Working Paper #4: Policy Incentives Diversification and Comparative Advantage. 12/ All whetit consumed in Sri Lanka is imported. 10 1.34 and 1.38.'3 Earlier EPC estimates (1990 Maha)'4 of between 1.23 to 1.47 show little or no change.5 Average EPCs for paddy and import competing vegetable crops (based on chilies) in 1993 were 1.36 and 1.70. Both EPC estimates are very close to NPC estimates for the respective year, implying that output price policy played a much more dominant role than input price policy in setting the price-based production incentives for farmers. 3.7 EPCs - Export Crops. Fruits, vegetables, foliage plants, medicinal herbs, etc are the main export crops of the non-plantation sector. The export earnings from fruit, vegetables and foliage plants experienced steady growth over the last decade. Among the other minor exports, cinnamon and raw tobacco experienced stable growth. Export of other crops fluctuated widely over the last decade. Under export orientation, policy incentives include a subsidy program for providing financial assistance from Export Development Board (EDB) for investment, income tax exemption, waiver of duties on intermediate goods import under the Board of Investment (BOI) incentive package. The targets of the incentive package are big commercial enterprises whose capital investment exceeds Rs. 10 million and who export about 70-90 percent of their output. EDB informs that fruits and vegetables are not included under this program, and national accounts data show no subsidy payment. The incentive package for minor exports included fertilizer subsidy, direct subsidy to the producers and a price support scheme for coffee, cloves, pepper, nutmeg and nutmeg oil. In 1993, the direct and fertilizer subsidy amounted about Rs.9.6 million. The subsidy paid under the price support scheme was about Rs. 16.8 million. The national accounts data also show that the export tax on MECs was about Rs.0. I million. On the whole, the total subsidy was less than a percent of value added in the minor export sector. Including fruits and vegetables, the EPC for the export crops during the 1990-93 period, is estimated (from national accounts data) to average 1.02 implying a virtually neutral policy regime. The overall EPC for the non-plantation crop sector is estimated to be 1.24 in 1993.16 3.8 ESCs - Nonplantation Crops. Although agriculture also receives considerable subsidies on nontraded factors of production, these do not enter into estimation of EPCs. Including these would make the actual level of effective protection to agriculture much greater than seen from the EPC ratio. Irrigation, research, and extension, for which cost recovery is zero, are a major source of assistance to crop agriculture, which, when taken into account, materially affect the picture. Industry does not receive a similar volume of subsidies. For irrigation, since no charges are collected, the irrigation subsidy amounts to 100% of the annual capital cost repayment, finance charges and O&M costs. In large-scale irrigation systems, the construction cost subsidy is equivalent to an annual payment of about Rs.50,000 per ha (about US$1,000), or 96% of the economic value added per ha for paddy. The O&M subsidy here is equivalent to an annual payment of about Rs. 1,350 per ha, or 3.5% of economic value added for paddy. For rehabilitation investment, the subsidy equates to an annual payment of Rs.7,000 to Rs. 10,500, and an O&M subsidy equivalent to an annual payment of Rs.600 to Rs.1,350, depending on whether the scheme is a minor or major one. This equates to 21% or 27%, and 2% or 4% respectively of value added for paddy in economic terms.'" These are not implicit costs, but are actually incurred through the Government budget allocations in support of irrigation schemes. In return for this, on average rice consumers (primarily taxpayers and the poor) have paid about 20% more for rice (currently 30%) than they would if imports had been liberalized. 3.9 Clearly agriculture receives far more favorable treatment than might be perceived to be the case from the EPC estimates arising from Government intervention in markets in outputs and traded inputs. Taken with the EPC estimate, they cause it to increase by these same amounts. For paddy produced on a recently constructed large-scale irrigation scheme, the EPC is effectively increased from 1.34 to an ESC of 2.30. For other import competing and export crops grown on irrigated land the protection effect is similar to rice protection. The magnitude is not as great because of the larger value added in these. Overall, the average increase in protection nation-wide to paddy is about 20% and 10% overall for the nonplantation sector. Clearly irrigation subsidies add significant increases in protection, especially for new large scale irrigation systems. 1 3/ The EPCs tend to be smaller in the major irrigation scheme areas, where average paddy yield is much higher than other areas, i.e. minor irrigationi schemes and rainfed areas. 14/ Edirisingihe N., Abeyratne F.. Somarathne W.G.. Wickramarachi P., and Tudwe P.l., (1991), 'Efficiency and Policy Incentives in Rice Productioln in Sri Lanka", ARTI/IFPRI. 15/ Aniual variatimnsls in yields and/or inputs could account for the 10% difference I W, The reintroduction of the fertilizer subsidy in late 1994 will increase the EPCs for all crops in 1995. 17! See WP#3, Tatble 5.5. I1 3.10 ESCs - Plantation Crops. The plantation sector has traditionally been heavily taxed. In recent years this policy approach changed with elimination of the export tax on the sector. The EPC is now is estimated to be around 1.0 for both 1990 and 1993 (Edwards 1993). However, two factors not accounted for in these estimates are the fmnancial subsidies afforded the plantation sector by Government and the resident excess labor force of the plantation corporations, i.e. a subsidy and a tax. The ESC for plantation crops including these elements is 1.03. As with manufacturing, plantation crops are also highly disprotected relative to irrigated agriculture. Agriculture versus Industry 3.11 Manufacturing Industry EPC. The most recent estimate for this is 1.7. From the relative values of the EPCs, agriculture appears less protected than the manufacturing sector. Although the gap has closed somewhat in recent years, the average level of protection given to industry appears quite high relative to agriculture. The industry EPC above is restricted in coverage to imports only, and is derived from the tariff schedule rather than comparison of actual prices; but tariff collection has been subject to widespread, ad hoc exemptions, thus reducing the effective well below the actual rate. Also, in the past 2 years, import tariffs have been restructured into 3 bands, with a maximum of 35 % and a minimum of 15%, further reducing protection to industry. As a result, this estimate is seen as a gross overestimate of the actual level of protection afforded manufacturing imports. More seriously, by excluding manufactured exports, which now make up about 60 % of manufacturing industry output, the average level of protection afforded manufacturing industry is even more overstated. With no direct support from Government, the EPC for these is taken to be 1.00, the weighted average of the two gives an EPC of 1.30. This compares to an overall average ESC for agriculture of 1.32. Bearing in mind that the average for agriculture hides tremendous variations, this indicates that overall there appears to be little difference in protection given to the two sectors. 3.12 The degree of protection of the various crops relative to manufacturing industry is shown in Table 3. 1, with the crops, differentiated by type of irrigation system. Average nonplantation crops protection (ESC 1.45) is above that for manufacturing (EPC 1.3) and particularly above export manufacturing (EPC 1.00). Relative to export manufacturing, nonplantation crop agriculture receives far more protection. Only rainfed export crops receive the same zero protection as manufacturing exports. The extent of protection increases with the size of the system and is inversely related to the age of the system. This simply says that the higher the per-acre investment costs the greater the subsidy transferred. Large systems cost more per acre than small, and new systems cost more per-acre than rehabilitated ones. An ESC of 1.03 is estimated for plantation crops. This gives a weighted overall ESC for agriculture of 1.32. Allowing for the upward bias in the import substitution manufacturing EPC (1.70), and the inevitable increase in crop EPCs in 1995 as a result of the reintroduction of the fertilizer subsidy, agriculture is undoubtedly protected relative to industry. But the aggregate average hides tremendous variations in protection within nonplantation agriculture arising from the very high costs of irrigation and the complete absence of any cost recovery. Irrigated crops receive a far greater level of protection relative to manufacturing industry than rainfed crops, so that paddy grown in Mahaweli receives 230% more protection than do manufacturing exports such as garments. Paddy grown in rainfed areas only receives 40% more protection. But, export vegetables average only 3% more, while export vegetables grown in rainfed areas receive the same protection as manufactured exports, i.e. zero. Agriculture versus Agriculture 3.13 Within the nonplantation crop sector, the relatively favorable treatment given to paddy and other import- competing crops compared to export crops, i.e. on average 47 % more protection for rice (117% more for rice in Mahaweli) and 69% on average for other import-competing crops, shows a strong anti-export bias in agricultural trade policy too. This estimate is not subject to the concerns mentioned in respect to industry. Although other factors, particularly land tenure policy, are a severe constraint to sector performance, until the trade policy playing field is levelled diversification will not occur and sector output will likely remain very inward oriented. 12 Table 3. 1: Sri Lanka Trade Policy Protection - Crop Agriculture Relative to Manufacturing Industry. EPC | EPC | ESC Crop Protection (ESC 1993) relative to: l | Import Substg. Export Manufacturing Crop 1990 1993 1993 Manufacturing. Manufacturing. Average. Rice - rainfed 1.47 1.34 1.34 0.79 1.40 1.03 Rice - in a new major irrigation system 1.39 1.34 2.30 1.35 2.30 1.77 Rice - in a rehabilitated major 1.23 1.36 1.63 0.96 1.61 1.25 irrigation system l Rice - in a rehabilitated mnior 1.31 1.38 1.60 0.94 1.60 1.23 irrigation system Rice - average 1.33 1.36 1.56 0.92 1.59 1.20 Import-Competing Vegetables - rainfed na 1.68 1.68 0.99 1.35 1.29 Import-Competing Vegetables - in a na 1.68 2.26 1.33 1.57 1.74 new major irrigation system Import-Competing Vegetables - in a na 1.70 1.82 1.07 1.41 1.40 rehabilitated major irriganion system l Import-Competing Vegetables - in a na 1.72 1.80 1.06 1.39 1.39 rehabilitated minor irrigation system Import-Competing Vegetables - average 1.28 1.70 1.79 1.05 1.39 1.38 Export Crops (Vegetables) - rainfed na 1.00 1.00 0.59 1.00 0.77 Export Crops - in a new major na 1.05 1.25 0.74 1.16 0.96 irrigation system l Export Crops (Vegetables) - in a na 1.03 1.10 0.65 1.04 0.85 rehabilitated major irrigation system l Export Crops (Vegetables) - in a na 1.01 1.05 0.62 1.03 0.81 rehabilitated minior irrigation system l Export Crops (Vegetables) - average 1.00 1.02 1.06 0.62 1.03 0.82 Nonplantationi Crop Sector Average na 1.35 1.45 0.85 1.28 1.12 Plantation (Tree) Crops 1.00 1.00 1.03 0.61 1.03 0.79 Agriculture Average na 1.24 1.32 0.78 1.20 1.02 Import Substituting Manufacturing. na 1.70 1.70 . 1.70 1.31 Export Manuacttiurine na 1.00 1.00 0.59 - 0.77 Manufacturing Average 1.80 1.30 1.30 0.77 1.21 - Sources: Nonplantation Sector 1990: - Edirisinghe N., Abeyratne F., Somarathne W.G., Wickramarachi P., and Tudwe P.I., (1991), 'Efficiency alId Policy Incentives in Rice Production in Sri Lanka", ARTIIIFPRI. Nouiplantation Sector 1993: Para 3.8 and Working Paper #2. Averaged using: Within Crops area wts. Rainfed 0.31, New 0.04, Maj. Ir 0.41, Min Ir. 0).24. Between crops VA wts Rice 0.30, Imp. Comp. 0.33, & Exp. Comp. 0.37. Plantation: Adjusts public sector EPC (1.0) for financial subsidies and labor taxes and weights with smallholder plantation EPC (1.0). Agriculture Average: Weighted average, using 1993 value added shares as weights. Nonplantation 0.68, Plantation 0.32. Import Manufacturing: - Edwards C. (1993) Protectionism and Trade Policy in Manufacturing and Agriculture: Sri Lanka Institute of Policy Studies. Estimation process gives upward bias, but used in the absence of any meaningful basis to adjust for this." Export Manufacturing: - aLssumes no export subsidies and refund of import taxes on inputs. Manufacturing Average: - weighted average of industrial production, 60% exports and 40 % import substitutes. IS/ Edwards also estimated NPCs and EPCs for agriculture. The NPCs are similar to those here, but the EPCs are seen as too high due to omission oif finanice charges. See Working Paper #2. 13 Commarative Advanta2e - DRCs 3.14 The efficiency of resource use in production is measured by estimating the Domestic Resource Cost (DRC).19 By valuing inputs at their respective shadow prices, the DRC, takes account of the taxes and subsidies applied to the non-traded inputs also. For paddy in Sri Lanka, the DRC ratio varies from 2.16 in new largescale irrigation systems to 1.53 in rainfed production.20 It is generally higher in regions with higher yields, implying a negative relationship between yield and DRC, due to higher irrigation investment and O&M costs. But the estimates exceed unity for all regions, and the average DRC is estimated to be 1.73 (Maha 1993). This clearly indicates that more than one Rupee of resources (Rs. 1.73 in fact) is used to produce one Rupee's worth of rice valued in foreign exchange. The policy implications here are serious, as this shows that Sri Lanka does not currently possess comparative advantage for one of its major crops and that rice could have been obtained for less intemationally. This problem may have worsened in recent years as the average DRC, estimated under similar assumptions for Maha 1990 was only 1.22.21 The DRC for chili also exceeds unity, being 1.68 in Anuradhapura (rainfed), and 2.17 in Jaffna (irrigated), implying that, on average, the economy also suffers losses in production of these. All estimates take a 6 year construction period and a cropping intensity of 2. Because these values are over-optimistic, the DRCs are biased downwards, making the situation look better than is likely to be the case.22 Clearly this has lessons for future irrigation investment in Sri Lanka, a fact ably demonstrated in a recent International Irrigation Management Institution (IIMI) publication.22 3.15 Exchange Rates. Exchange rate overvaluation is often claimed to be a reason for such high levels of protection and lack of comparative advantage (i.e. EPCs and DRCs). In the past, the exchange rate in Sri Lanka may have been overvalued, thus taxing tradables such as rice. This does not appear to be so at present. With the DRCs for the large- scale irrigation projects reaching towards 3, throwing all the adjustment on the exchange rate would require a 200% depreciation of the present exchange rate to provide comparative advantage for rice produced in Mahaweli. But the exchange rate regime has not been the source of any substantial bias against agriculture. Partial equilibrium analysis, using varying scenarios, indicates little if any exchange rate overvaluation.24 Consequently, no significant adjustment in the level of protection, i.e. sectoral EPCs, appears likely on this account. Irrization Subsidy - Cost and Distribution 3.16 Developing and sustaining this production required substantial Govemment expenditure, primarily on irrigation infrastructure. These are almost all transfers from taxpayers to paddy producers.2" With no cost recovery, the size of the total transfer here may be gauged by multiplying the subsidy per ton of crop in the various systems by the volume of output. Based on 1993 output levels, this amounts to about Rs. 12 billion (US$240 million) for paddy and Rs.400 19/ The DRC ratio shows the cost of domestic resources used to produce a unit of foreign exchange earning. Alternatively, it shows the value (cost) of the resources used to produce a unit of the product (e.g. rice). The more above I the greater the loss involved in production, while die more below I the greater the economic profit and comparative advantage. 20/ In Sri Lanlika. irrigarion water is provided free of any cost. The Operation and Maintenance (O & M) cost per year is estimated to be about Rs. 1556 per ha in rehabilitated major irrigation schemes outside Mahaweli area, followed by Rs 1331 per ha in newly constructed major irrigation schemes ii Mahaweli area and Rs.600 in rehabilitated minor irrigation schemes. Using market wage and exchianige rates, the DRC ratios are estimared under the assumption that irrigation cost is equal to 0 & M cost in rehabilitated schemes. 21/ Edirisinghe N.. Abeyratne F., Somarathne W.G., Wickramarachi P., and Tudwe P.l., (1991), 'Efficiency and Policy Incentives in Rice Production in Sri Lanka", ARTI/IFPRI. 22/ See WP#2, para 5.7 (page 15). 23/ Aluwihare. P.B. and Masao Kikuchi. 'Irrigation Investment Trends in Sri Lanka: New Construction and Beyond.' International Irrigation Management Institute, Colombo, Sri Lanka, 1991. 24/ WP#3, Annex I and Table Al.3, Estimation of Equilibrium Exchange Rate 1993. 25/ Because die price elasticities of supply and demand for paddy are still low, the deadweight element of these costs is quite small and iistead it is a transfer to paddy producers. The elasticities are currendy estimated as: -0.67 for demand and 0.09 for supply. See WP#3. 14 million (US$8 million) for chili, or a total annual cost of about US$250 million.26 The per ton loss, 25% of import price ot paddy and 20% for chili, is significant. At the national level, this amounts to an annual transfer of about 3% of GDP in 1993. 3.17 Farmer families resettled in large-scale irrigation schemes were selected from congested, low-income areas. This must call into question the merits of large-scale irrigation investment as an approach to poverty alleviation. Rice self-sufficiency is the alternative argument, especially where farm families are no longer poor. But the very high and onigoinig cost of this approach to self-sufficiency must question its sustainability as well as its wisdom. If the sector had to pay its own way, as the overall economy must, then paddy would need an enormous increase in protection against imports to allow it to meet this bill. This would simply have the result of shifting the subsidy burden from Sri Lankan taxpayers to Sri Lankan consumers. 3.18 The massive irrigation subsidy mainly benefits farmers in major irrigation systems and so its distribution is extremely skewed. In fact, the subsidy per ha to farmers in recently constructed major irrigation systems amounts to 5 times that given to major rehabilitated major systems and 7 times that for rehabilitated minor systems. For rainfed systems, the ratio is infinite, as these receive no irrigation subsidy. But only 4% of the cropped area is in new large scale irrigation systems, while 41 % is in major rehabilitated schemes. Minor schemes account for 24% of the cropped area; while rainfed area is 31 %. Farm size appears to be much smaller (rural congestion much greater) outside the major irrigation systems, where settler activity is subject to greater administrative control. This adds to the skewness of the distribution of the irrigation subsidy. If'informal division of holdings in Government's irrigated settlement schemes is taken into account, this may not be as significant. Irri2atlon Policy Reform 3.19 Full correction in the current situation requires a policy of full cost recovery (capital costs as well as O&M) from beneficiaries of the irrigation system, i.e. farmers; as only this will shut off the excess demand for underpriced irrigation land. But this will still only ensure financial viability. To move towards a fully economically efficient irrigation system requires a system which permits market forces to determine the price of irrigation water through interaction of supply and demand for this. Because irrigation water is a private good, establishment of a market system is quite feasible, although, because the water is expensive to move around (including allowing for third-party effects), markets are likely to remain very localized. 3.20 Perceived market failures in irrigation systems appear the underlying reason for Governments' traditional treatment of it as a public good.2' These perceptions arise mainly from the existence of major scale economies in storage and distribution structures and from the absence (or non-enforcement) of water rights.2' Externalities, mainly due to the private cost being less than the social cost for many competing water uses may also cause market failure. Natural monopolies of large-scale storage and distribution schemes give rise to market imperfections only. 3.21 Treating irrigation water as a public good inhibits much needed institutional reform, and the creation of independent and (eventually) private irrigation agencies. Policy makers and planners continually call for better performance by public-sector irrigation institutions, despite the absence of institutional autonomy and adequate financial incentives for personnel. The situation is well summarized by Repetto: 'For the most part, management problems are symptomatic of the underlying conflicts in the political economy of public irrigation. But, many of the remedial projects deal mostly with the symptoms and not the underlying conflicts. If performance in public irrigation is treated either as a mechanical or design problem, or as a management problem, and the more 26/ Inlsufficient data were available to estimate the DRC for onions and potatoes. It is seen as less than that for chili (due to higher transport costs), but the combined volume produced is much greater (5 times), suggesting economic losses for these on the order of perhaps another US$10 million. 27/ The key defining characteristics of a public good here are 'nonsubtractabilitv' (the ability to consume as much of a good as desired with no reduction in the aimount available to others) and 'nonexcludabilitv' (the inability to exclude specific individuals - e.g. users who fail to pay a share of the costs - from consuming a good). Street lighting is a good example of both. 2h/ For a detailed discussion see: Rogers, Peter. Comprehensive Water Resources Management: A Concept Paper. World Bank Working PLaper WPS 879, Matrchl 1992. 15 fundamental difficulties of the political economy of public irrigation are not resolved, efforts to improve performance will probably have limited success. `29 3.22 Capturing the scale economies in storage and distribution requires massive investment, for which Government has traditionally been the source. This has strengthened Governments' view of irrigation as a public good, leading to an exclusive public sector administration for the system. The administered system is entirely "supply driven", in which the water users have little or no say. Administrative pricing of the water assigns a cost-plus value to water and ignores the underlying supply-demand situation. This generally results in undervaluing the water by comparison with what users value it at and encourages waste in use. To correct this, an approach needs to be developed which recognizes and addresses these vital shortcomings automatically and in an ongoing fashion. In other words, what is needed here is an institutional structure which would permit direction of system operations by market forces, and thus convert the system from one which is "supply driven" to one which is "demand driven. 3.23 To restore financial and physical sustainability and also efficiency to the irrigation system, hidden gains or rents need to be made clear and tradable; there needs to be a direct connection between water charges and deliveries. At present, there is no direct commercial relationship between irrigation water suppliers and users, and no management autonomy in irrigation agencies. Both are major factors in the breakdown of administrative, nonmarket, discipline. Options to address the problem are: (a) direct legal intervention to restore discipline, by exhortation and coercion; (b) revival of the existing agency through reorganization and realignment of staff and responsibilities; (c) turn the system over to user groups; (d) restructure the agency as an autonomous public utility, with a hard budget constraint and a mandate of self-sustainability; or (e) privatization. 3.24 Exhortation has no lasting impact, and coercion is unacceptable. Internal reorganization, regularly tried, regularly fails. Being unable to address the basic underlying problems, it can only address the manifestations of the problems. Often, there are unbridgeable differences in opinion and interests between the legislature and the irrigation agency. This results in little administrative independence for the irrigation agency, which anyway faces a multiplicity of objectives (many of which are neither commercial nor consistent). 3.25 The solution lies in commercializing any service or product which is not a public good. Irrigation is not a public good and so should be commercialized. In this way, market incentives can promote needed improvements in efficiency. Within each scheme, an autonomous public utility, independent of the regular civil service, is an absolute minimum. This approach can only be as successful as the Government's commitment to the independence and commercialism of the utilities. Complete privatization in the form of a user-owned and operated utility company is preferable. This removes any doubts about independence and commercialism, clearly resolves property rights issues and avoids the possibility of private investors exploiting farmers or the need for a regulatory authority. A judgement needs to be made in the case of each system as to whether or not it is ready for such an institutional turn-around. 3.26 A public utility will permit development of the needed two-way links between the supply and demand sides of the irrigation water market. Financial self-reliance will provide the stimulus for utility management to ensure cost- effectiveness and profitability. It will also provide the incentive to collect water-delivery charges and those for other services, thus restoring financial discipline in users. Users in turn will demand improvements in operations and will only pay for what they get. Farmer Organizations (groups of individual users), would be a counter balance to the monopoly power of the utilities, but only if adequately developed. Finally, legalization of water sales between users (individuals and groups) will permit efficient pricing of water in local markets, leading to more efficient on-farm water use. Linking drainage service charges to water delivery charges for collection by utilities could enhance drainage cost recovery, ensuring financial viability for drainage O&M and helping with finance for drainage expansion. 3.27 While the institutional and legal aspects of any restructured system should be flexible, the basic elements must be present. Creation of legally enforceable water property rights and legalization of markets in these (lease and sale) are vital to progress here. Creating user water rights and legalizing water trading can be expected to provide a transparent market value for water, and its opportunity cost of the factor. This promotes more efficient use of this scarce resource, through equalization of the marginal value product of water in its alternative uses. 29/ Repeulo, Rubert C. Deceniber 1986. Skimming the Water: Rcnt-seekinig and the Performance of Public Irrigation Systems. World Resources InStitulte, Rcsearch Report #4. Washington, D.C. 16 3.28 Institutional options for any irrigation scheme will vary depending on the social development of the local area. Where democratic and developed farmers' organizations exist, they should, together with other user groups, (local Governments, industry associations, and environmentalists, for instance) take over system operation. In practice, this is occurring at an increasing rate in a few countries--Spain, Chile, the US, and (most successfully) Mexico, for example. Although the Bank supports such user organizations, their development is of necessity slow. Even in Mexico user organizations must pass through a well-monitored development phase before the system is turned over to them. 3.29 The incentive for user groups to take over their system depends very much on the their ability to perceive and capture otherwise foregone benefits. The benefits are generally in the form of lower costs and an improved water supply. To the extent that farmers in Sri Lanka make informal payments (rents) to acquire 'free' irrigation water, the first possibility holds. But accessing this benefit in a cost-effective manner requires that user groups act as directors of the public utility administering the system; otherwise, the practice of extracting informal payments will continue. If timeliness and equity in delivery of water is a problem, again user groups will have an incentive to take over the system. Such a take-over is best phased in, as user groups are organized, federated and trained to carry the responsibility of operating their own system (e.g. as in Mexico). 3.30 Where such incentives are not available, an autonomous, self-financing, commercial public utility is required. In existing systems, full cost recovery (investment and O&M) may not be possible without excessive shrinkage of economic activity. Since investment costs are sunk, recovery may be limited to rehabilitation and O&M costs without hindering efficiency. If farm income per ha is unable to bear even rehabilitation costs, then recovery may be limited to O&M costs only, with no rehabilitation investment undertaken. In effect, this means running down the fixed asset, i.e. an eventual shrinkage in irrigated area. 3.31 With increased costs the quality of management input to farming becomes more important. As a result, it is likely that the number of farmers will decline as the less efficient ones move permanently off-farm and average farm size will increase. This will result in increased profitability in agriculture and offset forces tending to shrink irrigated area. The entire process is simply the normal result of price-cost squeezes, and one designed to ensure continued economic efficiency in fanming systems. It is likely that conditions in Sri Lanka's land market will hinder this adjustment process as it has up to now (see Chapter V. Land). Trade Policv Reform 3.32 The subsidy given to paddy prices directly arises from the import protection given to paddy and is borne by all domestic rice consumers in direct proportion to their volume of consumption. The relative burden depends upon the importance of rice in the purchases of the individual or household. Obviously, for the poor, any additional cost for rice will be a much heavier burden than for the wealthy. Measured by the NPC in 1993, the increase in rice prices over what it would have been in the absence of any import protection is about 30%. Taken as a share of the gross value of rice production, this amounts to about US$125 million, or 2% of GDP. 3.33 Because trade policy reform can bring rapid results, a more detailed analysis was made of the likely outcome of trade reform within the agricultural sector only. For this, a simple multimarket (partial equilibrium) mode130 was developed to reflect the relevant features of the sector. The model draws together actual sector data on variables of the major commodity groups"' in the form of a series of interactive markets which adjust to change via own and cross elasticities ,32 3.34 The impact of exogenous changes in policy variables on key elements of sector performance are assessed through the changed equilibrium levels of the markets included. The main results are presented below. Due to the partial-equilibrium nature of the model, not all the intersectoral impact of change is seen. As a result, the bulk of adjustment to policy change falls on the agricultural sector. But, as seen above, most of the policy-induced distortions appear emerge from within the sector. The transparency of the modef's approach makes the results intuitively acceptable, giving confidence to the direction, if not the precise magnitude, of the results. 31)/ For details of the multimarket analysis see WP# 3. 31/ Output, prices, income, employment, trade, and policy (tariffs, quotas) variables. 321/ The whelt flour subsidy, introduced in August 1994, is not seen as sustainable and so is not included in the model. 17 3.35 The policy reforms examined progressively extend partial liberalization of trade restrictions to the major output markets impacted by trade policy; beginning with rice, then incorporating import competing crops, i.e. chilies, potatoes and onions, and finally sugar. Reducing the rice import tariff by 10% impacts on both rural and urban households. Both groups are subdivided to represent three income groups, lower, middle and upper. 3.36 A 10% reduction in the rice import tariff causes a reduction in the cost of a major food, leading to a decline in the cost of living for both rural and urban populations. Real income increases for all, except the largest farmers (high income rural households), even though nominal farm income declines by 1.23% due to the decline in the rice price. This results from the price decline effect dominating the income effect, together with the fact that much household consumption is financed by off-farm earnings.33 3.37 The largest increase in real income goes to rural low income households (0.77% for the 10% tariff reduction), as these are almost entirely net consumers of rice and rice is the largest component of food purchases. This is 9 times greater than the gain to high income rural households. Incomes of urban groups at all levels increase, as these are all net consumers of rice. The next largest income increase goes to the low income urban households (0.71 %), for whom rice is a also a major part of household expenditure. Gains here are twice as large as those to high income urban households. This underlines the very negative effect of trade protection on rice for the poor, and so confirms the earlier findings of the Bank's study on poverty in Sri Lanka,4 and the very positive distributional impact Qf trade reform. Rice imports and consumption all increase, while production of import competing crops (chilies, potatoes and onions) declines, due to a substitution of rice in aggregate consumption. This also happens to tree crops, but is offset by increased exports. Production and export of high valued vegetables increases by 0. 1% and 6.0% respectively, offsetting the decline in production and increase in imports of rice and other import-competing crops, demonstrating the reality of the anti-export bias already noted agriculturaltrade policy. There is a small decline in on-farm employment, 0.07%, reflecting the inevitable adjustment out of agriculture. Also, a decline in Government revenues of 0.2% results from a decline in the price of rice. Overall, the outcome is seen to aid export development and poverty alleviation. 3.38 Expanding this to include a 10% increase in the import quota for import competing crops show similar results, but with increased magnitude. While nominal farm income declines by 1.84%, again this is offset by a decline in the cost of living such that real incomes increase to a greater extent than previously for all rural groups. As expected, urban household income increases more than before, with the increase again being almost twice as large for lower than upper income urban households. Production of rice and export vegetables increase marginally at the expense of the import competing crops.35 Again, imports of rice and import competing crops increase, while Government revenues decrease largely due to decrease in tariff revenues from imports of rice and sugar. As before, export of export crops increases, but still by less than 1%. Onfarm employment declines by a modest 0.12%. The wage rate declines by 1.15%, but this is in isolation of ongoing expansion in the off-farm labor market, which has been pushing up agricultural wage rates quite rapidly in recent years. 3.39 Partial liberalization of sugar along with rice and import competing crops results in the same 1.84% decline in nominal farm income. Real income changes are more positive (or less negative) due to the price effect of the decline in sugar price. Gains to rural low income households are about 5 times more than to the rural upper income households. Gains to the lower income urban households remain almost twice as large as to the upper income urban households. Sugar production declines, but by less than 1 %. Reallocation of the sugar area gives a slight increase in rice production. Production of import competing crops declines but not as much as before. Production and export of export vegetables increases, by 0.04% and 10% respectively. Onfarm employment declines by 0.12% and Government revenue from taxes and tariffs on these commodities by 2.29%. The latter reflects the size of sugar's contribution to overall tariff revenue in the sector. 3.40 Overall, these trade reform scenarios show that the positive aspects of policy reform are very encouraging and the downside impact is remarkably small. Under the present structure, even with full liberalization, expansion in export crops will be slow initially, primarily because of the current low level of output and also because the supply response 33/ See IFPRI Research Report of November 15, 1984. on "analyses of a series of issues which influence the performance of the Sri Lankan food stamp program". Table 39 - Value of Production and Share of Income by Landholding Size, (page 101). Research funded by the Office of Nutrition of AID under Grant DAN -1275-G-SS-2124-)0. 34/ World Bank Report No. 13431-CE, Sri Lanka Poverty Assessment, January 11, 1995. paras 2.79-2.88. 35/ This result is very sensitive to the cross price elasticity of supply between rice and imported crops. 18 has been policy constrained in the past.' The highly positive income effects on the poor, both rural and urban, are very welcome. The gains to low income urban dwellers and the rural landless, and thus the budget through reduction in transfer expenditures, provides a powerful reason to initiate the adjustment process. The anti-agriculture bias seen in overall trade policy must also militate against greater export oriented investment. But elimination of this has only a small impact initially, as the land market is nonfunctioning, and so potential investors would still have problems accessing the necessary fixed factor of production.3' However, the significance of off-farm employment on overall farm family income at present is very great (cf Chapter VI). Because of this, trade reform in isolation of other more fundamental reform may have a perverse effect on aggregate sectoral output.38 In this case, there may be a valid reason to delay trade policy reform pending adjustment in policies regarding land tenure and the rural credit market (cf Sequencing Policy reform, paras Price Fluctuations and Trade 3.41 It is sometimes argued that trade liberalization will cause greater fluctuation in prices of domestic food products, thus increasing risk to farmers and consumers and destabilizing domestic output levels due to the far higher level of fluctuation in world prices. A comparative analysis of seasonal and annual price fluctuations in the domestic and international rice markets shows quite the contrary in Sri Lanka.39 The coefficient of variation for monthly wholesale rice prices in Sri Lanka was 8. 1, while for rice priced fob Bangkok it was only 1.4 (both averaged over 8 years). The same measure of price fluctuation based on annual average price between 1984 and 1991 was 31.4 for Sri Lankan wholesale rice prices and only 15.6 for fob Bangkok rice prices. Clearly import liberalization could bring significant price stability to Sri Lanka's domestic rice market. IV. RURAL FINANCE Overview 4.1 As with most of the markets in Sri Lanka's agricultural sector, the rural credit market functions quite well. Competitively structured and with widespread retail coverage, physical access is available to almost everyone. Structurally, the sector shows increasing sophistication as different styles of intermediation evolve to carve out specialized niche markets for themselves. The smallholder structure of the nonplantation crop sector results in agricultural credit being restricted to a very small share of sectoral economic activity. In the event of liberalization and creation of conditions for growth in the sector, the structure and capacity of the rural financial system may need to be carefully monitored by the banking authorities. Mvlarket Structure 4.2 The supply side of Sri Lanka's rural credit market comprises three types of institutions: (i) formal, (ii) semiformal, i.e a hybrid of formal and informal, and (iii) informal.4' The formal institutions comprise banks which use the branch network to deliver banking services to the rural clients. Semiformal institutions are a hybrid system of both formal and informal institutions, which combine features of strength of both institutions and provide links between formal and informal systems. The semiformal institutions derive resources from within and outside the economy4' to cleliver credit and other financial services. The informal institutions are the traditional rural credit structures such as 3(1 It is likely that ltie supply elasticity for paddy is much lower than would be with a fully operational land market. 37/ The non tLUICiE1io 1m1g or water markers is likely to have a similar effect. 3X/ Depending on the divisibility ot the labor input in paddy production. 39i See WP#6. Table 9. 4(1/ For details see WP# 4. 41/ Menilbers jivhigs and grant financinig frotm oreign sources. 19 moneylenders who apparently use largely their own funds to lend to rural clients. Other informal lending includes own and extended family sources. 4.3 Formal sources are: the two state-owned bank's, Banik of Ceylon (BOC) and Peoples's Bank (PB) with 287 and 327 branches respectively, two-thirds of which are rurally located; the Rural Credit Department of the Central Bank of Sri Lanka (RCD/CBSL); and four private commercial bank's, the Commercial Bank of Ceylon (CBC) the Hatton National Bank (HNB), Sampath Bank, and Seylan Bank. Currently, these lend a total of about Rs. 10 billion annually, about 12 % of which goes to agriculture, including for both plantation and nonplantation crops, livestock and fisheries. 4.4 Semiformal sources comprise two Government controlled institutions, 17 Rural Regional Development Banks (RRDBs), and Cooperative Rural Banks (CRBs), and some 150 independent NGOs who engage in financial intermediation to varying degrees. The RRDBs have 160 branch offices. Although coordinated by the CBSL, which nominates each RRDB's 5 person board, they operate independently in pursuit of deepening the rural financial structure. The CRBs are offshoots of the Multipurpose Cooperative Societies and work closely with Peoples Bank in mobilization, lending and pawnbroking. The primary NGO is the Thrift and Credit Cooperative Society (TCCS) known in Sinhalese as SANASA, with about 731 thousand members in 7,632 village-based primary organizations, which are federated into a national level organization. The Sarvadaya Shramadana Movement (SSM), is the second largest. It differs from TCCS in being focussed mainly on the poor (those with difficulty in accessing credit). Presently, TCCS resources come mainly from obligatory member savings, which, with an interest spread permits own resource generation. It now generates about Rs.600 million annually and lends about Rs.700 million. With an overall loan/deposit ratio of 1:2, it is very much self reliant for funding. Most other NGOs, including SSM, aim primarily at nonconventionally credit- worthy persons, obtaining some resources from external grants and from the IDA financed Janasaviya Trust Fund (established 1990). The volume here is relatively small and not aimed at commercial agriculture. 4.5 Informal sources of credit are mostly village based and quite amorphous. They include full-time and part-time money lenders, traders and family lenders. With overlap and gradation across the spectrum, it is difficult to directly establish the contribution of the various sources. Credit extended by traders is likely to be prefinanced in part by the formal sector. CBSL reports indicate that a substantial amount of informal lending comes from local persons engaged in fulltime service activities (including public-sector employees). Own-financing maybe the most important source of input financing in the rural sector. 4.6 On the demand side, in addition to agriculture and fisheries, loans in rural areas are utilized for small-scale industry, housing, electrification and water supply, and "projects/commerce" (i.e. trade). These account for 60-65% of all advances from formal and semiformal Fls in the rural sector.42 Credit Financing 4.7 Crop production financing from the banking sector is provided by an array of formal sector FIs, but primarily the two public sector banks, Bank of Ceylon (BOC) and Peoples Bank (PB). Sources of funds are: i) own funds; and ii) own funds onlent under a Government guarantee (New Comprehensive Rural Credit Scheme). Market rates are charged for own funds, while NCRCS loans carry concessional rates.43 Although the specifics of coverage change, interventionist schemes of this nature have been a feature of Sri Lankan rural credit for the past 30 years. The rate of uptake has been low relative to input use. Two-thirds of NCRCS financed loans are for paddy production and the rest for import competing crops. 4.8 Term financing for agriculture, mainly for food (rice) processing, is provided by the banking industry, both public and private. Multilateral loans are a sizeable source of funds here, including the Bank's lines of industrial credit. Onlending rates are market oriented, being set at marked up deposit rates. 4.9 A comprehensive measure of the annual volume of short-term credit to the non-plantation crop sector from the formal and semi-formal Fls is elusive. In the meantime, NCRCS loans appear the best available measure. These have 42/ CBSL Annual Report 1993, Table 1.70. 43/ The NCRCS concessionial onlending rate is 16%, compared to market rates of around 23%. Government refunds the difference to the onlending Fis anid also guarantees to cover 50% of all loan losses. Up to 1994, the NCRCS loans were eligible for refinancing by the Centrall Banki of Sii Lalika's (CBSL). This is currently discontinued. 20 been unrestricted in volume and carry concessional lending terms, as well as protection for lenders in the form of a guarantee of 50% cover for all losses. These features make it very attractive to both lenders and borrowers, and unusual circumstances would be required to transact cultivation loans outside of the program. Since its inception in 1986 through 1993, NCRCS credit to the nonplantation sector, has increased by 19% p.a. or 6% p.a. in constant volume terms. By comparison, Purchased Inputs have only increased by 0.22% annually in constant volume terms over the same period. But the share of NCRCS financing of Purchased Inputs only increased from 5.3% to 7.9%. The rapid growth in NCRCS financing reflects the small base at the outset, as well as the likelihood that some RFIs substituted NCRCS financing for their own resources due to its relative attractiveness (cf para 4.19). Interest Rates, Savinigs and Input Financini 4.10 Interest rates under NCRCS loans are compared with commercial loan rates and savings rates below. The relevant deposit rate for small farmers with uncertain cash flows is the regular savings accounts rate (i.e. the minimum rate); the higher rates apply to time deposits and savings certificates. Over the entire 1986-94 period, despite the concessionary nature of NCRCS, the loan rate has well been above the deposit rate, making it far less costly to finance purchases from own funds than from NCRCS credits. At the lower end of their range, commercial bank rates appear competitive with NCRCS rates. But the weighted average prime rate of commercial banks, 14.2% p.a. in 1986 and rising to 20.5 in 1994, has always been well above the NCRCS rate. So, although small farmers do not qualify as prime borrowers, this of no loss to them. Table 4.1 SRI LANKA - Agricultural Loan and Deposit Rates - 1986-94. Period NCRCS COMMERCIAL BANK LOANS SAVINGS LOANS Immovable Unsecured DEPOSITS Collateral 1986-90 9% p.a. 12-30% p.a. 13-33% p.a. 6-14% p.a. 1990 12% p.a. 9-28% p.a. 13-35% p.a. 5-14% p.a. 1991-92 14-16% p.a. 9-30% p.a. 13-33% p.a. 6.5-14% p.a. 1993 16% p.a. 13.8-28% p.a. 16.5-33% p.a. 6.5-14% p.a. 16.5-28% p.a. l 1994 16% p.a. 15-28% p.a. 16.5-36% p.a. 6.5-14% p.a. Sources: NCRCS - CBSL Rural Credit Department. Others - CBSL Annual Report 1993 Table 89. 4.11 In 1992, a typical year, and the most recent for which comprehensive data are available, the record for paddy and OFCs was: Total Purchased Inputs Rs. 11,228.8 million44 Cultivation Loans - NCRCS Rs.864.9 million45 7.5% of Purchased Inputs At first glance, the low level of credit financing of inputs (7.5 %) might be seen as a constraint to sectoral performance. However, CBSL data for 1992 show Cultivation Loans as less than half of all loans and only 14% of all savings. 44/ National Accounts of SL 1993 - Final estimates Table 9 (p 63). Excludes hired labor, much of which is not a cash tmansaction. 45/ CBSL AniiiLutl Report 1993 Table 1.69. 21 Total Savings - Rural Sector Rs. 5,983.7 milliont Total Loans - Rural Sector Rs. 2,148.1 million4' 36% of Savings Cultivation Loans - NCRCS Rs. 864.9 millione 14% of Savings and 40% of Loans. In other words, over 90% of purchased inputs are financed from informal sources. Total Rural Savings net of Cultivation Loans equals about half the Value of Purchased Inputs, showing substantial surpluses available in aggregate after financing of input purchases. This indicates that, in aggregate, farmers can finance their purchased inputs from earnings prior to saving. 4.12 Cultivation Loans in 1992 account for about 30% of Total Loans in the Rural Sector and Crop Loans for only 30%. Two-thirds of all loans are for nonagricultural purposes, mainly housing, electrification and water supply. Rural dwellers seemn to prefer to borrow for housing improvement over agricultural inputs. This could be taken to indicate a supply-side credit constraint arising from formal and semiformal RFls, restricting own-funds loans for home-equity investments, due to the fact that the latter may act as their own collateral, whereas farm land is non-mortgageable, which would limit the total of input credit to the amount covered by Government schemes (NCRCS), However, Total Loans are only about 1/3 to 1/4 of Total Savings (36% in 1992), while Cultivation Loans are In7 of Total Savings (14% in 1992). Consequently, it is clearly feasible to expect most inputs to be own-financed, unless rural savers and farmers are not the same group, which in aggregate is unlikely. The negative spread between the deposit and loan rates provides a strong incentive for own financing of inputs. 4.13 Rural sector income appears quite sufficient for all input financing at present. Also, the informal market competes successfully with the formal market, even where this has access to concessional credit. One conclusion of interest to policy makers is that, in this situation of high levels of off-farm incomes, the deposit rate, not the borrowing rate, is likely to be a significant factor in determining the volume of inputs used in the nonplantation crop sector. Credit Constraints 4.14 It is often contended that there are nonmarket-based factors that constrain credit to agriculture and so limit sector performance. Structurally, the rural credit market would appear to be quite competitive. Entry and exit for FIs is easy; a sizeable number exist, and with a large branch network, making local monopolies unlikely. However, specific supply- and demand-side aspects of the credit market have been highlighted as evidence to the contrary. 4.15 On the supply side, the relatively low level of purchased inputs, high transaction costs, the development of semi- formal, group-help based RFIs, and the existence of a concessional credit program (NCRCS) and its lack of expansion, are all advanced as indicators of constrained credit supply. On the demand side, inability of potential borrowers to meet eligibility criteria is put forward as a constraint to expression of demand for credit. 4.16 Purchased inputs, at 25% of value added may appear low; however, to attribute this to a credit constraint only may be very misleading. Despite its lack of comparative advantage, Sri Lanlka's paddy yield is comparable to that for similar Asian countries, suggesting a surfeit of inputs if anything. Also, the opportunity cost with respect to off-farm employment (cf Chapter V) questions the value of additional onfarm effort in terms of income generation. These would be demand shifters, but not exogenous constraints to the supply of credit. 4.17 Transaction costs may be higher than necessary, due to the lack of managerial autonomy in public-sector banking. High administrative costs (heavily state-oriented ownership and restrictive labor regulations have lead to over- staffing). While policy determined, this is not aimed directly at credit supply. The same situation does not arise in private sector commercial banks, making a prima facie case for privatization of the two public sector commercial banks. Also, while this is one area capable of improvement, it does not appear to have had any impact on credit availability, only on the cost of credit. Also, this is offset by the concessionally priced NCRCS credit program. 46/ CBSL Annual Report 1993 Table 1.70. i.e. rit amount afrer financing purchased inputs. 47/ CBSL Annual Reporn 1993 Table 1.70. 4XI CBSL Annual Repon 1993 Table 1.69. Includes PB and BOC data. but may or may not includc sornc or all rmral loan data in Table 1.70. 22 4.18 Development of a semi-formal credit supply is seen aimed at expanding access to credit, and so income generation. Structurally it lies between formal and informal credit sources, competing with both. In competing with formal sources, it endeavors to overcome factors constraining access to formal suppliers (mainly collateral, biased selection, and lack of credit discipline.)'9 To do this, it uses group-discipline (supervision) to substitute for collateral (or heavy supervision) and enforce a sustainable credit culture. To compete with informal lending sources, mainly moneylenders, it maintains low overheads and prices it loans on a cost plus basis. In this fashion, it has successfully carved out a market niche. Its success appears to lie in its ability to supply credit at a lower cost, thus substituting for formal and informal sources rather than creating new credit. TCCS, the major player in semi-formal credit supply, has only 25 % of its annual loan portfolio in crop agriculture. Since TCCS is demand driven (by its members), this lending pattern is likely to be responsive to their demands rather than being supply driven. 4.19 Semi-formal RFIs rely heavily on own funds for resources, so the impact on the rural savings rate will be a reasonable indicator of whether or not the semi-formal sector is a credit creator or substitutes for other sources of credit. Unchanged rural savings rates implies credit substitution and increased rates, credit creation. The share of savings generated, while believed substantial, is indeterminable due to the unavailability of PB and BOC data on savings mobilized by source (rural and other). However, if, as CBSL indicates,"e rural savings mobilization is mainly by the semi-formal institutions, this implies no change in the savings rate induced by the expansion of semi-formal RFIs, and so no credit creation, but possible improved access to credit. 4.20 The concessional NCRCS credit program faces no policy determined supply constraint. Instead, the amount utilized is limited by demand for input financing and competition from other sources, mainly own savings financing. Discounted by the recovery rate, the effective average cost of NCRCS credit to farmers (16% x 0.7 = 11.2%) is above the deposit rate. At best this leaves farmers indifferent as to the use of their surplus earnings between saving and purchasing inputs. When even minor transaction costs are added to credit financing, the balance comes down against credit financing of inputs. Since concessional agricultural credit has been available for about 30 years in Sri Lanka, and branch banking is widespread, lack of information concerning availability is unlikely as a reason for non-expansion. The high default rate,5' makes all formal cultivation loans unprofitable; but FT losses under NCRCS are under half that of losses on own funds.52 As a result, private sector commercial banks are unwilling to enter this field, and.the public sector banks will restrict activities here to the extent that these losses are uncompensated for.5" The high spread and the high default rates certainly constrain NCRCS loan volume, but again these are endogenous constraints, and not attributable to any anti-concessional lending stance. 4.21 Difficulty in accessing credit due to an inability to meet eligibility conditions, often used to infer an exogenous constraint on ability to express demand, is more likely to be an endogenous supply-side effect. As seen above, high default rates, combined with limited loss guarantees, generally result in losses to Fls, and so a reluctance by formal sector FIs to enter or expand this credit activity.5' To limit their losses, FIs may try to select out high-risk borrowers. Although no more than risk-weighted pricing, this is often incorrectly seen as discriminating against subsistence farmers. At present, no private sector commercial banks undertake NCRCS lending." 49/ Collection rates for these semi-formal RFIs are generally close to 100%. But even for the semi-fornal RFIs with group based lending, collection rates for agricultural loans are well below there average. Consequendy, even for these RFls, loan portfolios are heavily biased against loans for purchased inputs in agriculture. 50/ CBSL Annual Report 1993 (page 165), Mobilization of Rural Savings. 51/ Averaging about 30% since 1970, it dropped to 159% in the mid-80s, but has since risen again. CBSL Annual Report 1993, Table 93. 52/ For commercial loans, Rs.(123xO.7)-123 = -Rs.36.9, i.e. Actual return minus expected return gives a loss of Rs.36.9 for every Rs. 100 lent; while for NCRCS loans, the loss is only Rs. (I1 6x0.7)-116 + ((116x0. 7)-116)x.05 = -Rs. 17.4, or 47% of the loss under comnercial conditions (assuming no difference in default rates). 53/ Faced with a high level of liquidity, formal Fis outflank the indiscipline problem through udlizing local taders as intenmediaries by seasonal-financing of inputs stocks. Traders can employ good selection criteria, as being local, they avoid the asymptotic information trap, and being private they are not susceptible to political pressure. Finally, they will have low collection costs. The extent of this is unknown. 54/ In Sri Lanka, the fact that only the public sector banks participate in NCRCS is seen as evidence of this. 55/ The impact on private sector commercial banks' portfolio performance of debt-forgiveness in public sector rural credit schemes is unknown. 23 4.22 Government has endeavored to create a more favorable environment for banks through guaranteeing 50% of losses incurred under NCRCS, and strengthening the framework for debt recovery; the political process, to which public sector banks are highly amenable, may have severely weakened the end product. Over time, client selection by noncommercial Fls has not permanently impacted on the default rate. This suggests that selection may be beyond the prerogative of management, and instead determined by elements extemal to the Fls unconcerned with institutional performance. In this case, NCRCS credit would only be available to persons with some political standing in rural areas, which would constrain access by poor farmers. 4.23 Inability to use land as collateral is sometimes seen as a substantive obstacle to access, but very likely only for the marginal borrower,5" i.e. for the subsistence farmer, who faces the greatest likelihood of nondeliberate default. Concessional credit programs aim to overcome this pricing problem. Here it appears that NCRCS terms are insufficiently concessionary to overcome the risk to Fls. But the more generous the concession package the greater the budgetary cost. The answer lies in restructuring the approach to one of very restricted targeting, heavily focussed on subsistence agriculture as the appropriate target group. However, while not a constraint now, the nonavailability of land as collateral is likely to be a major constraint to development of a long-term credit market, which will be essential for a competitive land market. 4.24 Other, unpublished, data57 indicate that lending in 1992 and 1993, by formal and semiformal institutions, for paddy and OFC production was about Rs.5 billion, or 45% of the Value of Purchased Inputs. This (unpublished) volume is 6 times the amount lent under the NCRCS concessional scheme, giving Rs.4 billion of nonconcessional loans. Since there are no policy restrictions on the volume of credit covered by the concessional program, it is difficult to understand why either a lender or borrower would choose to operate outside the comforts of the program. But even if true, because this still leaves the majority of inputs as own-financed, and surplus savings remain unchanged, this does not appear to change the fact that no, nonmarket-based, credit constraint is in evidence. Credit and Subsistence Aericuiture 4.25 The existence of net savings in rural areas alongside concessional credit is seen due to the varying net positions of individual farmers, i.e. some are net savers and some are net credit users." Combined with stagnant output performance, this could also be interpreted as an indication of constrained access to credit. But as shown in Chapter VII, expanding output under present conditions can well result in loss of income, making this interpretation unlikely. However, this is only true so long as off-farm employment is available. Where it is not, it is quite possible to have a situation where a farmer has labor surplus to the farms requirements and output is constrained through lack of input financing. With competitive financial and input markets, and no evidence of any aggregate input-financing constraint, this situation is likely to be structurally induced, arising from farm size being viewed by the credit market as too small to be viable as the sole source of family income (cf para 5.19). 4.26 About 30% or so of the farmers (about 1/2 million farmers) on the lower end of the farm size distribution5 are the farmers whose productivity will be constrained by access to input financing if they are unable to find off-farm employment and/or access to credit. But these are the people Binswanger and Rosenzweig wrote about in their note "Are Small Farmers too Small to be Efficient", addressing the problems of micro-sized farms in South India. Poor farmers, due to their asset scarcity, do not have the capacity to carry risk (i.e. bear intertemporal losses), and because the cost of insurance is so high,' the market, however perfect, is unable to provide them credit. It is unrealistic to expect the commercial credit market to provide inputs to such people, whose needs should be addressed in a different 56/ Unsecured loans carry the same interest rates as collaterized loans - Table 4. 1. 57/ Provided to the mission by CBSL Rural Credit Department. 58/ Some may also he due to the defaulter group, for whom the cost of borrowing (0%) is way below the opportunity cost (the savings deposit rate). 591 i.e. widt a farm size of less than 0.5 ac (0.2 ha) i.e. the average was about 0.2 ac (0.08 ha) in 1982. Sri Lanka Census of Agriculture 1982 General Report. Table 8B (p. 16). 60/ Crop insuranice is very costdy because of the covariant nature of the risks facing the pool of insured (i.e. each insured's crop suffer from the same natural hazards as the next), the high cost of the asymptotic information and moral hazard problems (policing and paying). As a result, crop insurance is almost never successful and should not be seen as a possible solution. 24 manner, i.e through programs targeted to subsistence agriculture rather than through programs for commercial agriculture, including commercial credit. Rural Credit Svstem - Outlook 4.27 Because the agricultural credit system is relatively shallow (as a proportion of sectoral economic activity), the question arises as to whether or not it would have the capacity to meet expanded credit demand in the event of an upsurge in sectoral output? It is likely that loan appraisal capacity and ability would prove to be at least one limiting factor, especially where new activities are involved. This could result in a significant increase in nonperforming loans and endanger the financial strength of the credit intermediaries. The public sector banks already have sizeable problems in this respect. The semiformal intermediaries, while appearing to be relatively free of such problems now, do not have the size and diversity of operations to permit them to bear substantial losses. Where the sector is also undergoing a fundamental structural adjustment, particularly in terms of farm size, this concern would be enhanced. Thus in the event of rapid growth in the nonplantation sector, additional monitoring of the health of the rural financial institutions and close supervision by the Central Bank would be needed. 4.28 The likely requirement by rural financial intermediaries that loans be fully collateralized by fixed assets, such as land would reduce this risk. This could enhance the adjustment process through helping the less efficient farmers to conclude where their best financial interests lie. Because such collateralization would require freehold land titling and establishing a land market, it will inevitably act as a severe brake on expansion of sectoral economic activity as well as biasing the gains towards those already with access to financing and cause the premature exiting of otherwise well qualified farmers. This points to the need to move ahead as rapidly as possible with a land title reform program and establishing the institutions necessary for land titling and registration as a prerequisite to a land market. 4.29 The role of Government in rural credit must also change from one of providing concessionary credit which undermines credit discipline to one of ensuring development of a competitively structured credit system. It must also play it's part in resolving credit discipline problems through cessation of all loan-forgiveness programs and provision of systems for rapidly and equitably resolving all valid loan contracts. 4.30 Under fulltime professional farming, substantial credit financing would be required for to meet input needs, especially where output is expanding rapidly. The current high level of the nonconcessional interest rate, around 23- 25%, could prove too costly for many activities, especially new crops with uncertain returns. This high rate is not a reflection of the functioning of the rural credit market, but of the overall macrobalances in the economy. Trying to resolve or even alleviate the problem through direct interventions in the rural credit market will only create further problems, as any concessional scheme facing a sector with high credit needs must inevitably be rationed, leading to biased distribution, resource misallocation, nonviable credit institutions and growing economic and financial losses. Instead, any high interest rate problem needs to be addressed at its source, i.e imbalances in public sector revenues and expenditures. Policv Conclusions 4.31 Overall the rural credit market appears to work well. Competition is certainly present in aggregate, with all sources of credit supply in competition within and among the major segments of supply and no evidence of any supply constraint is present. On the demand side, the high level of own financing appears to largely eliminate any commercial credit pricing issues. Although these will inevitably exist for asset poor and remotely located individuals and communities, this cannot be taken as an indication of a poorly performing market. Institutional performance may also be weak in some RFIs, which if corrected would lower transaction costs; again this is not a sign of a distressed market. Thus, while improvements are possible to enhance the cost-effectiveness and coverage of the rural credit system, there is no evidence that this in anyway constrains credit availability in aggregate and thus sector performance. 4.32 Although NCRCS is a relatively small share of input financing and a substitute for already available financing, it still represents a significant cost to taxpayers, amounting to about Rs. 190 million (US$4 million) in 1993. The major concerns with access to credit appear entirely on the supply-side, but are seen to be endogenous market driven factorsS and unrelated to market structure or policy-induced factors. On balance, the credit market would appear competitive zind unconstrained by any non-market based factors, and so is unlikely to be a constraint to sector performance. The 25 concessional NCRCS credit system may well be creating an excess of credit. Terminating the program would both eliminate the excess and reduce the budgetary burden. 4.33 External influence in agricultural-input loan repayment, mainly from political sources, has greatly reduced credit discipline among farmers, who tend to see these loans as grants. The resulting losses to the financial system has led an unwillingness on the part of the RFIs to actively pursue lending for agricultural inputs. This contraction in financial intermediation in rural areas will need to be reversed to permit growth. In short, to reverse this, political intervention in loan portfolio management in this area will need to cease. The only feasible way to achieve this is to establish a structure by which such intervention becomes ineffective. Autonomization/privatization of the entire banking system (similar to the approach recommended for irrigation) is the only feasible way to achieve this. 4.34 Rationalization of Government involvement in the agricultural credit market should center around separating all concessional credit programs into those which are strictly commercially oriented and those which are aimed at subsistence farmers. Commercial programs should be sold to commercial banks, and any public sector commercial RFIs should be restructured as necessary and privatized. Noncommercial activities, aimed at subsistence farmers should be tightly targeted and structured so as to minimize leakage into commercial areas. In effect, this would be a subsistence agriculture project, with the objective of making these farmers commercially viable. V. LAND Land Tenure Policv 5.1 Land tenure policy and legislation in Sri Lanka emerged from the recommendation of the Land Commission of 1927 for preservation of smallholder agriculture through holding Crown Land in trust for the whole community. Successive legislation, in the form of the Paddy Lands Acts of 1953 and 1958 and subsequent amendments, regulated tenant farming activities, with a heavy bias in favor of permanency for the incumbent tenant and fixing the maximum rent in terms of volume of output (15 bushels of paddy per acre). The Land Reform Law of 1970, and the 1975 Amendment to this, nationalized large estates including tree crops plantations and regulated the size of paddy farms nationwide, alienating excesses to the public sector for redistribution to smallholders as tenants in perpetuity. This has left a ceiling of on private, owner operated paddy land of 25 acres, while for tenant farming and in public sector irrigation schemes, the ceiling is 5 acres. 5.2 The Agricultural Productivity Law of 1972 provided for dispossession of agricultural lands not used produc- tively. Subsequently, the Agrarian Services Act of 1979 restricted the use of paddy land to the production of paddy only, but limited dispossession to only voiding the owners right to cultivate such lands. The 1991 Amendment to the Agrarian Services Act provided for appeal against dispossession orders and eliminated the restriction on the cropping of paddy land. 5.3 The primary objective of land policy has been to provide fixed assets for landless farmers in the form of land. With a rapid post-WWII spurt in population growth, and in the absence of corresponding growth in other sectors, pressure on land grew more intense. By the time of the Land Reform Law (1970), 75% of holdings were already below 5 acres, so that the redistributive impact of the Act was small. Other efforts to provide land to the landless focussed on finding idle public sector land for redistribution. Estimates of up to 2 million acres of rainfed agricultural land have been made, but substantiation is difficult. In any case, this land, lacking in water, is bound to be marginal, requiring much larger farm size to ensure viability. Farm Size 5.4 In line with longstanding policy objectives, land tenure policy is heavily focussed on poverty alleviation, farm size is generally small and farm numbers and size in Sri Lanka's nonplantation sector are distributed relatively evenly (Figure 5.1). The most recent agricultural census (1982) showed 91 % of all farms were of 2 ha (5 acres) or less. These occupied 60% of the nonplantation agricultural area, but almost 100% of the actual cultivated area. Irrigation- development has been dovetailed with this land tenure policy through creation of resettlement schemes in major irrigation investments, providing a maximum irrigated area of 2 acres, plus 1/2 acre of "highland" or nonirrigated land. Currently 72% of farms are 1 ha (2.5 acres) or less. But this data is 15 years old and does not reflect the current picture. 26 Continued population pressure, with no significant rural-urban migration, varying legislative codes, multiple user and inheritance rights, and protracted legal procedures, have all resulted in widespread informal multiple co-ownership and further fragmentation in privately owned land, resulting in continuous decrease in effective farm size over time. Despite legislative prohibition, the same forces have inevitably led to informal fragmentation in state-owned lands (irrigation settlements) and, although not well documented, all indications are that this continues to grow.6' F~~ure 51: t4onplantadon Seci~or .amNmesm ........ % cumulative ............... .. . 90 iSil000j0-:0-CT!:.0. .j;Xi: g itE: iEESt!0.-,0 iDigai. i.00000.F.!t ti-400- tl.:.e, -igig!000.g;-it-E-;X0-:ift:-0X,li--,440.t :0000..... -..,-g.. ....... | :l l;:l i; i m i:; iw' it;00:;l :;:} l: - ti ::. :"'--;E t t: iS00.isu0.5.100^0'Non,UM*U ' '.,-',S.'.'... . '...".....'' '.".'.'.'."" P - :! lig :;;SEt , tigl: iS00; C;0-0 .esmiL.fiL:0.00tl;0-00": H as t. %. . . .. . . ...i.s... tS' '-0- i-fff :-S:-;: ' b it' (':: f-'0:S ;::f; '7 o: :!0:.f:f;t'S; S : :''! i ff:ft, 2.fff:E :; f : fiS -;::iL f ; A: -:-.:. S i f . i..-.... ... . . . . . . . . . . .. . . ..-. . . . . . . . . . . . . . . -. . . . . . . . . . . . . . . ... .. . . . . . . . . . . . . . ...:- : - .- S. ...-: - -- .-.,.,- : .... ,.*. . ..... . .. . . . . . . .. . . . . . A A A7N.. S a S~~~~~~~~~~~~~~~...... . Sourc.i $jt Lute AuIwjhur.i 0mev. li..S....t ....(.t....w.b....8... . . . ...:. -:0 -::g -;... . . .. . .. i .00!....F-ifi:d '::: d. iA .: . ..:f:i ../ ... .. .0-|, .:!i .. ....t: Fte0:!:, ......":'V .. . .TS": ..iS i;;;.iS. ;SA :-,i , ;-; .iSSf-SA ;# - . ...... i .. . ...-...-.ff.,. . .- . . . ...0-004-0-t. 00:.5... . . .. . . . .. ... : igE.. li7 . .7 .:- g- -5 |.:.-.... E: . . ... ..g;;-7. g g... . g- A-gg: 75g 5.5 With0 84% of;ii :ig gi :0 00 the total, thet s itate; dom tgina::-0-tll$ies lan ow erhi in Sri Lanka eve in agrculur (66 excldin : .}.f.-.f.;.. f.E.. !.Zl!AS; .... -- :ir :W.;W ,# f.:: i: - : i.,;-gLi-g,.i-i! 70: i: t!:,.;i;E .; ! Wit. i; ,. .iE.00E, i l. # .442E....... .... ......... .. ..70;. inland waters). Agricultural land accounts for approximately 2.79 million ha or about 43% of the total land mass. Of this 63 % is owned by the state, but farmed by private fartners under varying tenure arrangements. The distribution of this is shown in Table 5. 1. Tenancy arrangements vary in degrees of restriction as to use and transfer, ranging from very restrictive LDO leases (99 year leases with unitary use, access and succession, and acceptance of mortgages restricted to cooperative societies and seizure or foreclosure for debt repayment disallowed), to 'Swarnabhoomi" (Golden Lands)62 land grants (permanent titles, but having a significant restriction in a requirement for prior permission before transfer and registration of title as an effort to prevent subdivision and multiplicity of ownership). 5.6 As a result, neither LDO (Land Development Ordinance - 0.82 million ha) lease nor Swarnabhoomi grant lands (0.1 million ha) are acceptable as mortgageable collateral by commercial banks. This fact qualifies them as less than purely private property, although Swarnabhoomi lands are classified as privately owned in Sri Lanka. Although ongoing since 1982, the rate of conversion of LDO use permits to Swarnabhoomi grants has been extremely slow. Only 6% of state owned agricultural land is under Swarnabhoomi grant title. A more general restriction on sales of Governmnent settlement land is that it can only be sold to the same class of farmer, i.e., a Government-settled peasant farmer holding not more than 1 ha, so this limits the sales possibilities. Since settlement holdings are constrained in size to I ha, this .......... . ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~. 61/ See WP# 5. 6V/ Recently chaniged to) Jayahhoomi; whichl has approximately the same meaning in English as Swarnabhoomi. 27 further limits the sales possibilities. Finally, since the land is not mortgageable and such small farmers are unlikely to have sufficient savings, effectively any formal market is eliminated here. Land Utilization and Cropping Intensity 5.7 Both land use and cropping intensity are low. In 1992, permanent crops,63 utilized about 1.0 million ha of agricultural land, leaving about 1.7 million ha of agricultural land for utilization by nonplantation crops, i.e., annual crops.' Actual area cultivated for these crops65 was 0.69 million ha, or 44% cropped in Maha, and 0.4 million ha, or 25% in Yala. These are very low utilization rates are attributed by IIMI to a shortage of irrigation water relative to land.& Availability of water may constrain overall land utilization, but would not explain cropping intensity, particularly in irrigated areas. Paddy cropping intensity has been as high as 140% in the past decade. But this is now much lower. The recent 3-year average was only 109%.67 While the explanation would appear to lie at a more fundamental level, it does appear that people in agriculture may be deliberately reducing their farming effort. Inevitably this must impact very negatively on irrigation investment returns. But, it can only be seen as an indicator rather than an explanation of the poor performance in the nonplantation crop sector. 5.8 Time series data show the aswedummized (irrigable) area to be increasing at a faster rate than the gross area cropped by paddy, causing the ratio of the two, i.e. the cropping intensity to decline. About 80-85% of the aswedummized area, or 540 thousand ha, is irrigated. Government owned, large and medium scale schemes account for about 55 % of this, the remainder being much older, community developed, village-level schemes. But, these large and medium schemes generally have sufficient water for both Maha and Yala crops. These also account for the more recent additions to the aswedummized area, suggesting that the decline in cropping intensity is more likely to be in the older, village-tank schemes (cf paras 6.2 - 6.3). Change in rainfall pattern or poor maintenance could cause this too happen; in which case, the expected response would be a substantive change in cropping pattern away from crops with high water demand and low value towards higher valued and less water intensive crops. But this is not seen. In any case, the loss to the economy through low land utilization and cropping intensities is undoubtedly significant. CropDin2 Pattern 5.9 Despite its low relative profitability, paddy is by far the dominant crop in the non-plantation crop sector. It occupies 77 % of nonplantation cropped land in Maha and 81 % in Yala. 6B Income from paddy (GDP per ha) is among the lowest of all major crop groups in Sri Lanka. In the nonplantation sector, on average, GDP per ha from the widely varied Other Food Crops (OFC) group is 7-times that of paddy, up from 4-times in 1980.4 These margins appear too large to be explained by the average-versus-marginal returns to fixed factors argument. Neither does limitation in market demand for OFCs explain the situation, as individual suppliers are too small for this to impact on their output decisions. 5.10 Nonetheless, despite its relatively low profitability, paddy dominates the cropping pattern, utilizing a cropped area 4-times greater than that for OFCs. Since farmers are not charged for water, its cost does not enter into GDP calculations. Yet, water is the constraining factor in Sri Lankan agriculture and paddy is a highly water-intensive crop. Imputed costing for water, even at its financial cost, would push the relative profitability ratio even more in favor of OFCs. In economic terms, the cost to the economy of such a cropping pattern is seen to be extremely high. The cost 3.3/ Comprising tea. rubber and coconut, fruit plantations, and spices (Minor Export Crops) 04/ Mainly Paddy and Other Food Crops (OFCs), including grains, pulses, fruits and vegetables. 6%5. In 1989-90, the last year in which all island data was published. 66/ 'Potential of Diversified Cropping in the Paddy Lands of Sri Lanka". Mimeo version page 17. IIMI, Colombo, December 1994. (i7 Average of 1988/89-90191 - Agricultural Statistics of Sri Lanka 1992 Table 5.5 68/ Based on 1989-9t crop data, but year-to-year variation is small. Due to water constraints, Yala cropped area is only 60% of the area cropped in Maha. 6)/ See WP#6. Taible 10t. Also note, this fatetor vanes between 3 and 7 depending on the crop, year, and data base. 28 3;;0 X03 -CATEG:ORYE }i f030 i:: ~ L cc i 00- t:-:Inlandgilt ,,'gWate s 14ggli;:. {,20 m*ln hi I: -t : 5-02::T:iiFu0resi st&: Re :ig-rv :s 2g .18 nuill -- h tt ii it3 3% - t-: : * iA whtLh..~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...... - - unde...r O lease QZ millf l. . . vi : : :Xs.0 .000 .................00-$ 0. - . .- . T r ee.C.. . ...P l.at io. . ... .... . .Ur.- b aa : ArElel 0;2 I0 mil hi 0;i0 ....- ... -:i.t wi iich ...i e G;I7 m i h- S .- ' . '- ~ To0ial Large fnJarid Waters, -ores.,0ts and Wild bfe Reserves arid Urban Areat'P-'$L $at.i.' iro. f. . w) Jt.emaindei derive ahme%tally i-q psutm x*ll;Mith try s Ag-ureXa to e LrelyPwatelyi Owne (to Col Imb th ari0 .~ne or i0wmIe -~ h 20^ is 1 - noly b0 Wiaa4e I( gg -.1 ti Tnru 5;0tis ;;la,nl* wll Tb*!; - r - .k ;rtu ocmTreE:rnps And;mor~ Saw Lanka 19 iage 41 r h rn a ade o-f-f} Prvll Owei l: n -. I : ~ t b : ::!Z. Y-rginU*Wil lad ertdfve,4.a inty-:-4 -W: -gr...-t-r.l lnd ar i,-.f 2; 5g - d ... hi.t .. ~ c lanS Envir i su,rvs Es e h s ad b fe-R.......i E: ml 1......... ....T. of therrigationsus idyalone(i.. te t r b s 3% of G (a 3 fro -th impor tEasriff whfSich : + is bon bytecnueidsaohr2 fGPt h ot(aa33) 5.11 It wuld pparthavtthsdmnancreofpdy,a lw: vaud, lo famicm-gnrtn comdiy s att ig % %it.Sk'tfigS~~~...... Et .:: . .. .... .. .....Id.lLWt.: 0:: source of poor performance in the nonplantation crop sector. However, low valued output and small farm size would appear a sure combination to encourage these less profitable farmers to exit agriculture so as to improve their income earning possibilities. This is a common phenomenon, and occurs in many countries. It is generally accomplished by selling the fixed assets, mainly land, and entering the nonfarm labor market. This allows average farm size to increase, thus increasing individual farm incomes to a level comparable with nonfarm incomes, even where overall sectoral income may be growing very slowly. Labor productivity will improve as the poor performers will be the first to exit. Because of the lack of a land market, this normal adjustment mechanism is not available to the sector. Larger farms will also have the ability to bear risk and so enhance the likelihood of diversification as tulltime farmers seek to expand incomet- earning opportunities. 29 Land Market 5.12. Over the 4-year period 1989-92, the average annual value of total registered land transfers in Sri Lanka was Rs.714 million.7t' The nature of this land and the area involved is unspecified. Capitalizing the income stream from paddy gives a land value of Rs.135 thousand per ha.'' Assuming (overgenerously) all land transferred to be agricultural, gives an annual sales volume of about 5300 ha, or 0.025% of the stock. At this rate, it would take 200 years to turnover (reallocate) all privately owned agricultural land and half a century to reallocate the entire stock (ignoring multiple sales). Although many transfers will be intergenerational, this is a very slow rate.' The fact that over 90% of all transactions are for "considerations" rather than mortgage financed may be an indication of this concentration on intergenerational transfers. But it could also result from difficulty in getting mortgage financing for asset transactions where title is so heavily disputed. Information on the market mechanisms and transaction costs are sketchy and anecdotal at best, due mainly to the narrowness of the market itself.' 5.13 It is important to note that even with needed policy reform in both trade and irrigation, the cropping pattem would still remain highly paddy oriented. Even the development of land markets and full-time farmers would not change this. As seen from the results of the simple model,'4 a 10% reduction in import constraints on rice and other import competing crops (chilies, onions, etc.) actually causes an increase of 0.18% in rice output. Even increasing the rice elasticity of supply 5 times, from the generally accepted 0.09 to 0.45, only causes a 0.9% decline in rice output. As a result, it seems clear that paddy will remain very much the dominant crop. Policy Conclusions 5.14 In brief the position here is: - Government owns the majority of the agricultural and irrigated land; - farm size is very small, and declining; - cropping intensity is low and declining; - the cropping pattern continues to be dominated by low valued paddy; - the formal land market is unable to contribute to changing farm size; but - needed policy changes will not bring radical cropping changes; paddy will still dominate. 5.15 The net result is that expected adjustment is not occurring. Normal adjustment in farm size is blocked by the combination of Government dominance of land ownership and an ineffective land market. Policy makers should look to amending this situation as it is a hindrance to adjustment and so sector performance. Farmers with nonviable holdings and low productivity farmers are being trapped in agriculture without the possibility of exiting short of abandoning their fixed assets. 5.16 A generalized policy of private ownership of agricultural land should be adopted. Accompanying this should be programs in settlement areas to transfer ownership to listed tenants without delay. In private sector land, mechanisms to rapidly resolve disputed title should be established. Swift action by all concerned parties (public and private) should be encouraged through a tax-based incentive system to speed up both title transfer (public sector) and dispute resolution (private sector).'5 Programs should also be developed to enhance the efficiency and transparency of the land market, including providing at cost, all public sector generated and stored information concerning land records, policies, 70/ For details see WP#6, Table 7. 71/ Assuming a cropping ihtensity of I. 1. For details see WP#6. Table 11. 72/ Normally a finmn generation lasts about 35 years (i.e. from age 25 to 60); thus all farm land would change ownership at least once in this time period. 73/ For a description of land laws, public secEor land institutions, and sample indicators of the extent of title disputes, see "Land Tenure Study', Mimeo. S. Berugoda, Colombo, December 1994. 74/ See WP#3, Table 6 and Appendix i. 75/ The rationale for this would be compensation of the state for lack of full economic use of agricultural land. 30 programs, and transfer processes. All restrictions on minimum size holdings should be eliminated'6 to legalize the status quo and permit more orderly, market-based reallocation. Restrictions on land use, beyond those based on demonstrably valid urban zoning and environmental considerations, should all be eliminated. There appears no strong rationale for an upper bound on farm size, and this should be increased to not less than 500 acres to permit development of commercial agriculture. 5.17 The financial value placed on irrigated land will be critically dependent on cost recovery policy for public sector expenditures on irrigation investment and O&M. Continuation of the current zero cost recovery policy will result in a value much higher than with a full cost recovery policy. To avoid confusion and needless financial stress, Government will need to be completely transparent and consistent about its irrigation cost recovery policy from the start. This will allow the market to discount the land value for this, even if timing is uncertain. Inconsistency and abrupt change will cause undue financial stress and may lead to exploitation of some group or other of land owners. Also, consideration should be given to restricting sales of newly privatized Government land, for perhaps 3-5 years, to prevent unwanted sales gluts at the outset of the program. However, leasing of all types should be legalized immediately. Demand-based titling programs for public sector land might assist here, but run the risk of being designed to bureaucratically delay the privatization process. 5.18 The institutional elements for land markets need to be established. This will take time and require legislative action. Title Registration requires passing a "Registration of Title Act" to establish the basis for registration, reconciliation of disputes, transfer of title, and administration of the process. Allied to this is the need for a "Surveying and Mapping Act" to provide for the operation land surveying activities, including minimum qualifications, registration, licensing and bonding of professionals, distinction between public and private sector activities. Training needs capacity for public and private sector agents should be provided for. Because of the critical limiting nature of water in respect to productivity and value of agricultural land in Sri Lanka, unless its property status is established at the outset, uncertainty here will depress the value of land. As a result, creation of private water property rights will also become necessary to create the flexibility necessary for efficient functioning of factor markets. Drafts of both acts were approved in principle by the Cabinet in 1990. These, and other land related legislation, should be reviewed to ensure focus on facilitation of development of an efficient market system and contain no obstacles to this. Sri Lankan officials should establish an operating timetable for this. VI. FARM FAMILY INCOME AND OUTPUT ADJUSTMENT 6.1 The issues uncovered in respect to the incentive structure, rural credit, and land markets do not explain why there is no adjustment in the cropping pattern from a high-cost low-valued paddy to a lower-cost higher-valued mixture. These issues are external to the farm; instead, the explanation for this must be sought within the farm. 6.2 Rural areas contain about 65 % of the population and 40 % of the labor force, but agriculture accounts for only 25% of total GDP. With an average farm of I ha and 0.81% irrigated, a cropping intensity of 1.1, and cropped by paddy, average labor utilization is only 106 days (120x0.8xl.l), and yields an annual net income of about Rs.25 thousand, or Rs.5000 ($100) per family member. This compares with a national average of Rs.28,395 or $588 per capita (1993). Left to depend on agriculture, rural dwellers would be far less well off than their urban cousins. Clearly, the income differential is too large to be sustained in an economy with an open labor market, particularly since the average scenario utilizes less than one third of the available work-time. In fact, a substantial volume of income is earned off-farm. Labor Force Survey data for 1993 (see Table 6.1) 7 show that persons whose main occupation is agriculture spent on average one third of their working time in a non-agriculture occupation, and derived 41% of their income from non-agricultural work. For some, this can be as high as 80% of their income. Even persons whose principal occupations are non-agricultural average half their time working in the sector, although earning only 16% of their income in the sector. 76/ The public doesn't want them and they are unenforceable. 77/ For details see WP#6. Tables I and 2A. 31 Table 6.1: Sri Lanka - Labor Allocation and Returns by Occupation. % Non | |___________ J Agriculture | Share agriculture Share (Principal Occupation - Agriculture) _ Hours (million) 12.5 66 6.3 34 Inc. (Rs. million) 134 59 92.7 41 Wages (Rs./hr) 11 - 15 | (Principal Occupation - Nonagriculture) Hours (million) 24.2 52 22.8 48 Inc. (Rs. million) 158 16 818 84 Wages (Rs./hr) 7 - 36 - Source: Sri Lanka Quarterly Labour Survey 1993 - unpublished data. 6.3 Detailed investigations of this area are only now just beginning in Sri Lanka. An initial report 7, based on a survey of two villages (one wet and one dry zone village), confirms the aggregate picture seen from the Labor Force Survey data above. The villages averaged 200 (216 and 192) families and 4.75 (4.6 and 4.9) people per family. Farm size varied, averaging 0.30 ha aswedummized and 0.31 ha upland in the wet zone village, and 0.9 ha azwedummized and 0.30 ha upland in the dry zone.79 In the wet zone village, 40% of surveyed farmers and 90% of all households reported nonagricultural activities, while in the dry zone village 10% of farmers and 56% of households did so.' Non- farm income averaged 81 % of total family income in the wet zone village and 55 % in the dry zone village.8' 1 6.4 Analysis of family farm-income from the average farming situation in major and minor irrigation schemes and rainfed agriculture shows imputed hourly earnings rates ranging from Rs. I I to Rs. 14.' These compare well with the aggregate figure of Rs. 1 1 per hour for farm labor obtained from Labor Force Survey data. This Survey also shows, for persons whose primary occupation is agriculture, an off-farm average wage rate of Rs. 15 per hour. On the face of it, there might appear to be little advantage to working on or off farm, particularly for those at the higher end of the hourly earnings scale. But a more interesting fact emerging from the analysis is that the amount of family labor absorbed on the average farm is quite small, ranging from a high of 350 days on farms in major irrigation schemes to a low of 180 days on rainfed farms. At 1.5, the average cropping intensity'M in the analysis is well above the actual 78/ Report on Workshop on Rural Diversification (Project No IDRC 90-0117) May 27-29, 1994, Negombo, Sri Lanka. Socio-Economics and Planning Cemtre. Department of Agriculture, Peradeniya, Sri Lanka. 79/ Note this dry zone village is in the Mahaweli irrigation scheme area, but the farm size is well below the minimum authorized area of I ha. 80/ This and the Labor Force Survey, confirm the preliminary findings in regard to increasing off-farm employment in Sri Lanka Poverty Assessment, World Bank Report No. 13431-CE, January 11, 1995, para 1.22. 81/ For details by type of employment see Working Paper #6, Table 1. 82/ With no discernible change in either area sown or method of sowing for paddy (broadcasting 80%, transplanting et al 20%). paddy output levels clearly did not suffer froilm expanded off-farm employment opportunities. Also, it appears that increase in aswedummized area offset the impact of the decline in cropping intensity. 83/ These are averaged over paddy and major OFCs and over both Maha and Yala seasons. Disaggregation by crop and season widens the range to Rs. I to Rs.23 per hour. 84/ Assuming major and minor irrigation schemes as state owned land and rainfed as privately owned gives a weighted cropping intensity of 1.5. 32 average of 1.1. Also, the farm size used (I ha) may overstate the actual average. Consequently, the estimated the rate of family labor utilization may be on the high side. The average rural family comprises 5.1 persons, providing 1.965 years of labor, or about 590 days.5 The overarching conclusion is that, even where fully farmed, the average Sri Lankan farm remains a source of surplus labor, inevitably requiring family members to seek off-farm employment to meet demands for higher income. 6.5 If the cropping pattern were adjusted to 100% OFCs (chili, onion and potato) and 0% paddy, family farm income and family labor utilization would increase. Income and family labor use respectively would increase by factor of 2.4 and 3 for major irrigated farms, 3 and 2.8 for minor irrigated famis and 3 and 2.4 for rainfed farms. In fact, the average family could not meet this incremental labor demand internally. But importantly, before the inevitable decline in output prices, on major irrigated farms hourly eamings would decline from Rs.14 to Rs.12 per hour (15%), giving greater incentive for off-farm employment. Meanwhile, on minor irrigated and rainfed areas earnings per hour would increase from Rs. 11 and Rs. 14 to Rs. 12 and 18, or about 9% and 29% respectively. The increased hourly eamings on minor irrigated farms is still not quite competitive with off-farm possibilities (Rs.15), but the increase in rainfed areas clearly is. Yet here, as seen from village survey data (para 2.38), the level of off-farm employment appears to be higher than elsewhere. Three factors may account for the apparent contradiction. Wet zone (rainfed) areas are closer to major urban centers with larger labor markets and higher wages. Wet zone land, being largely privately owned land, suffers from far more fragmentation and greater title dispute, in many areas is climatically unsuitable for production of high valued OFCs, and, due to water shortages has low cropping intensities. This would result in reduced average farm size and income per ha, and so reduce hourly earnings to levels lower than estimated here. These factors are not well reflected in the summary analysis, but are clearly evidenced in the village survey study (para 2.37).86 This may also apply to smaller minor irrigation schemes, many of which are in the wet zone. In conclusion, it appears that the combination of high hourly retums to paddy and competitive off-farm employment possibilities sets the labor utilization pattern and farm output in Sri Lanka's nonplantation crop sector in favor of paddy (despite its relatively low return per ha) and off-farm employment.87 6.6 Within this scenario, it might be seen as possible to achieve some growth through increasing paddy yield. However, movement in the factors determining the trade off facing farm families here does not favor more intensive paddy cultivation. Most yield increasing efforts require significant incremental labor and management input, as well as more water and fertilizer.88 But real wage rates are increasing as is the fertilizer/paddy price ratio. Indivisibility may also be a factor at work here. Off-farm employmnent will likely be lumpy, as employers will require minimum volumes of time from employees. Farm activities will also be somewhat indivisible. Both will prevent family labor from achieving precise efficiency in allocation of their labor, and the balance appears tilted against spending additional time on the farm. In sum, real returns to paddy are declining, making it unlikely that anything short of a new technological breakthrough might reverse this stagnation. 6.7 The absence of an active land market may be at the heart of the low supply elasticities for most crops. Unable to adjust farm size to meet income needs, farm families seek alternative means to meet these needs. As shown above, the constrained optimum appears to be a mixture of farm and nonfarm activities. Unable to obtain sufficient fixed factors to match available family labor in an optimal fashion, and unable to exit agriculture without abandoning their fixed asset, farm families are constrained to be part-time farmers only. As a result, the cropping pattern is as much determined by off-farm as on-farm opportunities, causing a reduction in the normally anticipated response. But note that even with substantive policy change to bring the needed sector revitalization, paddy will still remain the principal crop (para 5.12). 85/ Farrmer I year. spouse 0.5 years, 3.1 children with 0.15 years each and 300 days per farmer per year available for work. 86/ Report on Workshop on Rural Diversification (Project No IDRC 90-0117) May 27-29, 1994, Negombo, Sri Lanka. Socio-Economics and Planning Centre. Department of Agriculture, Peradeniya, Sri Lanka. 87/ For furnher confinnation of this see Section 3.2 (pp 24-25) in: Dunham, David. Contract Farming and Export Horticulture: Can Agribusiness Revitalize the Peasant Sector in Sri Lanka?, Institute of Policy Studies, Research Studies Agricultural Policy Series No. 3. Colombo, February 1995. 88/ Rice Yield and Production Stagnancy in Sri Lanka. S. Wirasinghe and S. Emityagoda, Department of Agriculture, Technology Transfer Division, MADR, Peradeniya. Mimeo. 33 VII. TECHNOLOGY GENERATION AND TRANSFER Overview 7.1 The focus of attention of research and extension, which has been dominated by the public sector, has been on achieving rice self-sufficiency, a goal which has been very largely achieved. Since the mid-1980's, however, rice yields have stagnated, and there has been a growing realization that the research and extension system has not moved on to address the more pressing issues constraining production of higher value crop and livestock products, in response to market signals. However, there is only so much that the public sector can do efficiently in a dynamic market driven context where the private sector is likely to have a comparative advantage. 7.2 Research. Agricultural research has been dominated by public sector research institutes. Since the January 1994, restructuring of the Department of Agriculture (DOA) into autonomous commodity based research and development institutes, the Rice Breeding Research Center at Batalagoda has been known as the Rice Research and Development Institute (RRDI), and has been assigned the totality of work on rice. The Dry Zone Research Center at Maha Illuppalama has been given responsibility for other field crops consisting of condiments (chilies, onions), coarse grains (maize, millet, sorghum), grain legumes (green gram, black gram, cowpea) and oil seeds (sesame, soyabean, groundnut), and designated the Field Crops Research and Development Institute. The Central Agricultural Research Institute (CARI) has been assigned responsibility for fruit crops, vegetables, roots and tuber crops and ornamental plants, and has come to be designated as the Horticulture Research and Development Institute (HORDI). 7.3 Comparing agricultural research before 1977 with that since, there is a perception that output is low, its direction questionable and the impact on the farmer is small. However, private sector research initiatives have been very limited. Also, within the research institutes, there is no system whereby good research is duly rewarded, while remuneration and conditions of service (remoteness, etc.) are relatively poor. The Agricultural Research Project (ARP) has tried to help prioritize research expenditures through the introduction of Advanced Research Program Planning and a contract research program, together with improved staff training and facilities. But this has not remedied the overall sense of lack of direction which affects many researchers. Under these circumstances, the research system is widely perceived as supply driven. 7.4 Extension. The extension system has been disrupted by institutional changes. In the mid-1980s, after it was realized that agriculmral extension was costing the Government almost 1 % of agricultural GDP, with staff salaries alone accounting for 80-90% of operating costs, expenditures for the system were drastically cut. The Government also transferred 2,400 village-level extension agents (KVS) to its Poverty Alleviation Program. In addition, as part of the devolution of authority from the center to the provinces in the late 1980s, the Provincial Councils acquired operational responsibility for extension. Under the Second Agricultural Extension Project (SAEP), the Government has been trying to deliver better service to farmers at less cost. SAEP has fewer, but better trained, agents and makes much wider use of mass media. It has also established the farmer-centered farming system approach and, within this framework, introduced the use of problem census/problem solving techniques. However, many of the Provincial Councils still do not have the necessary capabilities or resources to sustain an effective system. Research/Extension Link 7.5 Despite the efforts of SAEP to strengthen the link between research and extension through the problem census/problem solving approach, research is still the responsibility of one provincial organization, while extension is that of another. Today's limited connections between research and extension are personal and informal. Policy Conclusions and Proposed Actions 7.6 Agricultural research and extension can only have meaning in the context of a well defined agricultural development policy, but this has been lacking in recent years which has led to an "identity crisis' in the research/extension community. In the new context of market driven development, however, the bulk of new research and extension, particularly in relation to higher-value export crops, is likely to come from the private sector, ask entrepreneurs supply inputs and technology to farmers as part of contract farming arrangements. What remains in the public sector needs to be made more demand driven. Government will still have a major role in extension, but a different one. Systematic provision of information on new technologies, farming systems, market prospects and prices 34 in a broad-based manner and in a form readily utilizable by farmers is needed. Since the externalities here are significant and the benefits from increased productivity substantial, this is a legitimate role for Government and one which needs to be recognized. However, services involving customizing such information in the form of designing farm management plans and recommending inputs for specific situations are better carried out by private sector extension services. Since much agricultural technology contains public good elements, this can make it difficult for private firms to appropriate the benefits of research. Consequently, Government will likely continue to play its traditional role here, pending more complete development of intellectual property rights. An overall assessment of the economic efficiency of the current and alternative models for generation and extension of technology which explicitly focuses to the above principals and considerations would appear to be a priority in this area. VIII. PUTTING IT ALL TOGETHER The Current Situation - What have We Learned. 8.1 Sector Performance is stagnant, with low cropping intensity, even on fully irrigated lands and heavy concentration on low valued paddy. Product Markets are competitive, but their efficiency would be improved if Government phased out its direct involvement. Input Markets are also competitive and work well. Trade Policy is inward looking, aimed at protecting domestic producers,resulting in non-competitive output. The Land Market is heavily constrained by property rights issues, including unresolved ownership claims in private sector land and an inability to reallocate land leased by the public sector and so permit minimum farm size to be market determined. The Rural Credit Market, although not cost-effective, is competitive and largely unaffected by Government's concessionary credit program for agriculture, which addresses poverty with concessionary credit, failing on both fronts. Overall Incentive Policy has a pro-agriculture but an anti-export bias, thus depressing potential export-oriented investment and so output in the sector. The Rural Labor Market appears to work quite well, as it substantially enhances farm incomes. But being highly informal, little else is known about it in terms of its ability to clear itself. More generally, regulation of the formal labor market must impact negatively on employment creation and on the efficiency of the informal market in the overlap area. Technology Transfer is a troubled area, lacking any mechanism for self-direction. Its excessive commnodity orientation limits its contribution to sector performance. Here, an in-depth economic assessment is badly needed to provide a basis for future direction and institutional structure. Policy Choices - Future Directions for the Sector. 8.2 In this situation, what are the policy options available to Government which might promote the objective of increasing sectoral growth? 8.3 The product and input markets are quite competitive; but their cost-effectiveness could be improved (marketing margins lowered) by elimination of all commodity-based intervention activities by public sector agencies. Also, additional efforts to promote transparency through well-run market information and analyses programs should show worthwhile improvement in functioning of the market. Amending the incentive structure to permit market determined investment flows to agriculture and remove the anti-export bias within agriculture would give substantial gains. Also 35 here, tarrification of all nontariff barriers should be accomplished without delay. However, as the multimarket analysis shows, these alone will not resolve the problem. 8.4 Access to credit is not a problem for commercial agriculture, but is for subsistence farmers. With substantial off-farm earnings and low growth, growth in demand for on-farm credit is inevitably slow. With enhanced sectoral economic activity tWis would change. In the meantime, the issue of institutional inefficiency and high transaction costs should be addressed. 8.5 Technology transfer appears quite confused as to its appropriate role and how best to organize itself for this. Insofar as commercial agriculture is concerned, privatization should be considered, leaving a public sector component to focus on areas where, for various reasons, the private sector may be reluctant to pursue. Programs specifically designed to meet, and focussed exclusively, on the problems of subsistence farmers need to be developed independent of those for commercial agriculture. Access to input financing and technology will inevitably be two important program areas here. 8.6 Although the rural labor market may appear to work well, as noted in Sri Lanka Poverty Assessment," part of the overall Sri Lankan labor market, it is constrained by Government regulation. Experience in India and Zimbabwe shows anti-retrenchment regulations depress job volume.' Thus, the recomrnmendation of the Sri Lanka Poverty Assessment report, to eliminate such regulation, should be implemented. 8.7 The inactive land market is a major problem. Because of this, farmers who wish to acquire more land cannot do so with any certainty in permanence of possession, which impacts on efficiency of use; while farmers who wish to exit agriculture are unable to do so in more than a part-time fashion, which also impacts on efficiency of use. By impeding the development of full-time farming, this in turn constrains the cropping pattern to one which severely constrains returns to fixed factors, primarily land, which is has been developed at very high cost, thus giving the economy very poor, and even negative, returns for large scale investment in this area. The remedy here is to initiate a land market. The Growth Challen2e 8.8 Within the sector, in 1993, Plantation Crops (tea, rubber and coconut) accounted for 24% of agricultural GDP, Paddy for 24% and Other 52%. Over the 1984-92 period, average annual growth9' in sectoral GDP was: Total 2.04%, Plantation Crops 0.16%, Paddy 0.22%, and Other92 4.08% p.a. If more rapid sector growth is the goal, 4.0% p.a. could be seen as a modest growth-oriented target growth rate for Total sectoral GDP. If Other continues to grow by 3.5 % p.a., and the combined growth rate of Plantation Crops can be stabilized at 0.5 % p.a., output from Paddy" must still grow by 7.5% p.a. to achieve the sectoral target rate. This is an unlikely achievement at present. Reducing the sectoral target growth rate to 3.5% and increasing the Plantation Crops growth rate to 1.0% p.a. still leaves a required growth rate from paddy (irrigated) land of 4.92% p.a., or 22 times the present growth rate. No amount of growth promoting effort for paddy itself is likely to achieve this. 8.9 This gives perspective to the dimensions of the challenge facing Sri Lankan agricultural sector policy planners and administrators. Since production of paddy and import competing crops already incurs major losses for the economy (para 4.8), this is clearly not an advisable direction for growth promotion. Instead, the entire nonplantation crop sector needs to be restructured to place it on an economically sound footing. 89/ Sri Lanka Poverty Assessment, World Bank Report No. 13431-CE, January 11, 1995 90/ Fallon and Lucas, The Impact of Job Security Regulations in India and Zimbabwe, World Bank Economic Review. Vol 5, No 3 pp.395-413. 91/ Using on 3-year averages at the beginning and end points. 92/ Made up principally of potatoes, chilies, onions. Also included are Minor Export Crops and Livestock. Minor Export Crops and Livestock each account for only 4% of sectorai GDP and so do not impact greatly on overall sectoral performance. 93/ i.e. the land occupied hy Paddy. 36 Reform Priorities 8.10 Public sector marketing activities, incentive policy, land markets, and subsistence agriculture are all priority areas to be addressed. Government should undertake a program to phase out of all its involvements in marketing, including privatization of CWE and PMB, and terminate all commercial activities of the Food Commissioner's Department. Incentive policy reform should aim at removal of the pro-agriculture and anti-export biases through a program of tariffication of all nontariff barriers and a phased reduction of these tariffs to a nominal level of 10-15%, i.e. just enough to dampen extreme external price fluctuations and retain some revenue for Government. The effect of price fluctuation on farm income, already small due to the size of off-farm earnings in total income, will be much reduced as farm size increases, and with it diversification and larger agricultural incomes. Since the impact on onfarm income is quite small, the gains to poor consumers attractive, trade reform should begin without delay. 8.11 Land tenure policy needs to be totally overhauled to facilitate development of a fully functional land market. This includes active promulgation and training programs on policies, professionals and processes related to sale and purchase of agricultural land, active encouragement of transactions through elimination of non-economically justified restrictions on ownership and size of farms, incentive-based policies to speed up privatization of public sector land holdings and speedy resolution of ownership disputes in the private sector. Current draft legislation needs to be reviewed to establish its adequacy here, and resubmitted to cabinet and the legislature for approval. Since land without irrigation water is clearly of little value in much of Sri Lanka, resolution of water property rights will also be necessary. 8.12 Due to the size of the problem, there appears an urgent need to upgrade the approach to dealing with subsistence farmers. While this is a dynamic problem and amenable to market solutions, past policy, which restricted the functioning of rural markets important to resolution of this, has constrained the adjustment process, leading to a ballooning in the current size of the problem. This group94 cannot be classified as commercial farmers due to the small size of their farms, or as landless farmers eligible for poverty alleviation type assistance. But they appear to be caught in the middle with no assistance program appropriate to their needs available to them. Left on their own, they are likely to be exploited as agriculture becomes increasingly commercial, as it must in order to generate returns commensurate with the resources made available to it. 8.13 Not all subsistence farmers will wish to, or be capable of, remaining in agriculture. Here the program should assist them to exit agriculture on the best possible financial terms, i.e. obtaining full market value for their land and training them for alternative off-farm employment. Others, who are potentially capable of becoming viable, should be taken into a custom designed extension program which would teach them technical skills and management practices on an intensive basis which would allow them to enhance their income generating capacities. Between the two extremes are those who are potentially viable farmers wish to remain in agriculture but whose farm size is too small. Development of land markets and suitable commercial credit facilities will take too long to assist these. Government could consider forming a land bank, using land acquired from those exiting agriculture to augment farm sizes of this group. To ensure viability, it would be necessary to run such a land bank on strictly commercial terms. 8.14 To encourage off-farm employment creation, rural employers should be exempt from restrictive labor legislation, particularly the Termination of Workers Act (TEWA). Employees should have the enforceable right, but without obligation, to form employee unions. Relations between employers and employees should be determined by direct negotiation between representatives of both parties. The role of the state here should be to ensure adherence to agreed bargaining rules. Concerns for basic health and safety should be addressed separately under independent legislation. 8.15 Other areas in need of attention by policy makers are credit and technology. The general approach in credit should be to end all Government intervention in commercial credit activities and confine its interests to provision of adequate liquidity to the system and ensuring its financial soundness and structural competitiveness. This should be accomplished through the Central Bank and the Ministry of Finance. In this vein, all public sector rural credit intermediaries should be privatized by selling them off to the private sector and closing down those for whom there are no buyers. '4/ Based on the 1982 Census of Agriculture, a total of between 0.5 and 1.0 million farmers and their dependents, or 2.5 to 5.0 miilion people are estimated to be in the nonplantation sector. 37 Seauencine of Poliev Reform 8.16 Correct sequencing is important so as to maintain momentum and avoid conflicting policy situations and undesired results. The major elements of the reform process which are seen to be substantively interdependent, which including all markets and policies which impact on sectoral income, and are in need of adjustment to improve their functioning, need careful examination in this regard. Thus, the main fixed factor market, land, input markets for irrigation water and long-term credit, and the import markets for competing products - rice and fruits and vegetables, are all of concern here. There may also be a need to introduce taxation measures into the sector to provide a more balanced investment climate between assets in this sector and in others. 8.17 Land and Credit Markets: Development of land markets is a major prerequisite for a more dynamic, growth- oriented nonplantation sector. But in the current situation, pursuit of this alone is unlikely to make much progress. Without liquidity in the form of long-term credit, farmers are unlikely to have financing available to purchase land. A major impact of this would be an inability of farmer-buyers to compete with investors. It will also result in depressed land prices, and so a reluctance on the part of sellers to participate. Avoiding this requires that deficiencies in the rural credit market, namely collateral to secure long-term loans and much improved credit discipline. Collateral requires marketable land titles. These would go to improving discipline, along with cessation by Government of all loan forgiveness programs, and assistance to the credit market in speedily resolving nonrepayment of claims. These interrelationships clearly illustrate the important symbiosis between credit and land markets. 8.18 Import Protection and Irrigation Cost Recovery Policies: Reform in here will reduce profitability of rice. If pursued in isolation, this may increase output of other higher valued crops, and so returns to fixed factors, i.e. land and irrigation. Since these are more labor-intensive crops, farmers will need to spend more of their working hours on-farm than at present. In the present situation, this will only occur if off-farm wage rates decline relative to hourly returns from these other crops. But this implies a retrenchment in the non-farm sectors of the economy; something which is certainly not desired. Greater on-farm labor input by individual farmers would also occur if farm-size could expand to where farmers could specialize in full-time farming and earn incomes at least as good as what they can now earn through the combination of paddy farming and off-farm employment. This requires that land can be reallocated at its full value, i.e through a transparent, formal land market. Otherwise, reduction in rice profitability from policy reform here may cause farmers to put more work time' in off-farm employment and less into farming, reducing rice production and actually depressing output in agriculture. 8.19 Irrigation Cost Recovery Policy and Land and Credit Markets: Future policy here is also very important to the functioning of the land market in Government-owned irrigation systems. The value of the land is significantly affected by its access to water and the cost of collection, storage, and delivery of the water to the land. The present policy is to charge zero for this service. This is (or will be) reflected by the market in the financial value of the land, and any future cost recovery here will reduce this market value of the land. Cost recovery here is not determined by any market functions; instead it depends entirely on Govemment policy. As such, it is not in anyway responsive to market conditions and so is uncertain. The real concem here is uncertainty not cost recovery. This uncertainty impacts negatively on the process of market-valuation of land. The market will significantly discount land values to allow for policy change in cost recovery which negatively impacts on farm income. Often, the discount will be overestimated to err on the side of caution by buyers and this will depress the market. This uncertainty will be magnified in the credit market. As a risk reduction practice, banks will generally extend long-term loans in amounts less than the 100 percent of the asset collateralizing the loan. Here, with the land purchased as collateral, the liquidity available to the land market will be substantially decreased, further depressing the land market. 8.20 Policy altematives available to Government to reduce the cost to the taxpayer of providing free water to irrigated agriculture range from the politically difficult policy of charging and collecting water fees to transferring responsibility for each system to the users, to transferring ownership of the system to the users. Each altemative requires a decision as to what charges to collect as part of the policy, i.e. delivery charges incorporating both capital costs and annual O&M costs, or, in the case of system transfer, charges to recover capital plus the capitalised O&M costs already incurred. Again, any uncertainty regarding Government's long-term policy stance here will be discounted in the market value of land and again in the collateral value assigned to it by the credit markets. 95/ Depending on the variability of labor inputs in paddy production. 38 8.21 Agricultural Taxation: Small-scale agriculture in Sri Lanka, as in many other developing countries, is not subject to direct taxation. In this situation, it is to be expected that development of a land market could well lead to significant acquisition of land by investors as part of portfolio management. To ensure that the focus of investment in land is for production rather than portfolio management, it may be necessary to consider introduction of agricultural taxation in the form of a transaction tax and resource rent or an imputed income tax on land holdings above some maximum size. The objective here should not be to exclude investors from purchasing land. This would only reduce liquidity in the land market and penalize those already holding land assets and wishing to dispose of them, such as farmers wishing to exit agriculture. The appropriate objective here is to avoid excessive investment in land as a store of assets, which would occur if the costs of acquiring and holding land were significantly less than the same costs for other assets in the economy. Such a situation would cause the market value of the asset to increase well above the present value of the earnings stream, making it difficult for professional farmers to acquire land. 8.22 A Pilot Approach: Improving the functioning of the rural credit market is a prerequisite and would appear relatively uncomplicated. Also, agricultural transfer programs should be separated into commercial and subsistence programs. Done simultaneously, this will enhance the acceptability of both. Land titling and registration will inevitably be slow. But it could be implemented more rapidly on an area basis, which, combined with advances in credit and commercialization of agriculture, could form the nucleus of a pilot program to field test the aggregate reforms. Clearly it is unlikely that trade policy could be packaged with this, but reform in current irrigation policy could and should. this could be based on either development of user property rights or straightforward gradual cost recovery. The property rights approach is preferable in the long run, and even in the short run when Sri Lanka's past history of cost recovery effort is considered. Government's Stance on Sectoral Reform 8.23 In its first major policy statement,' the current Government acknowledged the failure of past policies as demonstrated by "the present sorry state of agriculture and the rural population" and undertook to "remove the institutional and policy obstacles which have reduced agricultural profitability and constrained investment". The speech acknowledged that this would involve "substantial structural change", but that Government had "confidence in the rural masses and their capacity to make rational choices". Major policy changes in the nonplantation crop sector would include: - removal of hidden anti-agriculture discrimination from "inequitable macroeconomic and trade policies"; - eliminating all kinds of monopolies in agricultural markets and fostering greater competition; - realigning the roles of the public and private sector, particularly in irrigation; - creation of off-farm employment as the only viable approach to rural poverty; - improving rural infrastructure to enhance private investment in agroindustry and promote outgrower systems; and - freeing farmers to make the best use of their assets, including lifting restrictions on leasing land and granting freehold title in settlement schemes. In Governments' own words: "the new strategy is designed to ensure that the creative energy and independent spirit of farmers is not stifled by bureaucracy and over regulation." 8.24 The remarkable degree of alignment between the recommendations emerging from this study and Government's statement of policy intent indicates a firm grasp of the nature of the situation by Government. With the exception of agricultural markets, where Government rather than structure or conduct is the problem, the study's findings are in complete agreement with Government's policy intent. It is hoped that the study will provide greater depth to understanding the current situation and thus a basis for improved design of reform policies and measures. W6/ Policy Statement of die Government of Sri Lanka on the Occasion of the opening of Parliament, 6 January, 1995, by H.E. Chandrika Bandaranaike Kumara.unga. President. 39 Discussion and Dissemination 8.25 A brief visit in August 1995, introduced the report to the Ministries of Finance and Planning, Agriculture Lands and Forestry, and Irrigation (including MASL), CBSL, and the Institute of Policy Studies. Dissemination and extensive discussions were held from September 18 to October 4, 1995. A series of mini-workshops (9), each on different aspects of the study, was the approach. It allowed individuals and agencies, from Government, academia, and the private sector to focus on aspects of interest to them, permitting smaller audiences and more focussed discussion. Written comments and questions were also given to the mission by various sources. Summaries of the individual discussion sessions and responses to the major points of the written submissions have been circulated to GOSL officials and agencies. 8.26 Overall: Govenunent economic policy planners at the highest level are well aware of the issues and broadly share the Bank's strategy view, although they may differ on the detail of sequencing and timing in policy change. Reservations were expressed in the line ministries and agencies (Agriculture Lands and Forestry, and Irrigation - including MASL), primarily by noneconomists, about the findings overall. Much of this centered around the role of the off-farm labor market in explaining farmers' behavior, particularly with the combination of off-farm employment possibilities and land tenure policy motivating the production decisions of peasant farmers. But the ability of the integrated nature of the storyline to demonstrate, in an intuitively acceptable manner, just how the sector would adjust to the various hypothetical situations posed by discussants, including wholesale rural-urban migration, large-scale international conflict, and total stagnation in the nonfarm sectors, aided the non-economist audiences in appreciating the possible contribution of the study's findings in explaining how current sectoral performance is brought about. The size of the overall share of off-farm earnings in total farm family income in (42% in 1993) raised both scepticism and concern. Discussion of the supporting data 97 9S provided some perspective on this concern. 8.27 Background Papers: Six background working papers were distributed at the workshops. These provide the basis for the analysis and conclusions. It was strongly recommended that these be reviewed for their adequacy in substantiating the conclusions, and that the associated data and calculations be checked carefully for their accuracy. 8.28 Follow-utp: Direct follow-up action was not seen as possible at this time by MALF. Agreement was reached with MALF and the National Planning Department of the Ministry of Finance and Planning (NPD), that consideration should be given to several follow-up studies aimed at operationalizing the major findings of the report. i . Rural Wholesale Market Activities - a survey of volumes and costs to establish an information base; 2. Commercial Infrastructure in Large-scale Irrigation Schemes - the adequacy of this in facilitating rural commerce; 3. Land Tenure Status in Government owned Lands - i.e the impact of the informal market and how best to resolve negative effects, including needed legal and institutional changes; 4. Reducing Subsistence Agriculture - i.e., development of meaningful criteria to identify subsistence farmers, determine the economic factors primarily responsible for their situation, and propose means to resolve these; and 5. Pettah and surrounding markets in Colombo - the impact of barriers to entry to Pettah on development of spill- over markets and conditions of entry to these, and the impact of these developments on the overall efficiency of the marketing process. 8.29 It was agreed that the mission would prepare draft terms of reference for these studies for comment by MALF and NPD. Thereafter, these would be finalized for implementation under the associated Japan Grant for project preparation. 97/ Department of Census and Statistics, Ministry of Finance and Planning. GOSL. Unpublished data extracted from the Sri Lanka Quarterly Labor Survey. 1993. 981 The agricultural wage rate, which has risen rapidly, and the demise in production of high valued gherkins, are good examples of independent data which support the findings of the Quarterly Labor Survey. '4~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~' S .4 , s * ' s gk a . a e > ^ - -.; IMAGING R eprt No: 1,464 CE T y p e;