.a)_ ~E N V I R O N M E N T D E P A R T M E N T PAPERS PAPER NO. 15 _ "j~ .1J TOWARD ENVIRONMENTALLY AND SOCIALLY SUSTAINABLE DEVELOPMENT Lqt l. T ENVIRONMENTAL ECONOMICS SERIES Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh Stefano Pagiola May 1995 Environmentally Sustainable Development The World Bank ESD Land, Water, and Natural Habitats Division Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh Stefano Pagiola May 1995 Published in collaboration with Agriculture and.Natural Resource Operations Division South Asia Region Papers in this series are not formal publications of the World Bank They are circulated to encourage thought and discussion. The use and citation of this paper should take this into account. The views expressed are those of the author and should not be attributed to the World Bank. Foreword This report is part of a series of studies being Series since the issues it discusses are of much undertaken on Bangladesh's agricultural sector to broader interest than to Bangladesh alone. It is take stock of the current situation, identify pro- now widely recognized that sustainable develop- blems that need to be resolved, and establish ment is closely linked to a healthy environment. priorities for future action. It focuses on an issue The nature and magnitude of those linkages, which has beenthe source of considerable concern however, are often unclear. Studies such as this in recent years: environmental problems that one are necessary to establish whether, and in might arise from the growth of intensive agri- what way, economic activity might be affecting culture. These are issues that are critical for the the environment and to establish priorities for future development of Bangladesh: intensification action. As this report shows, agricultural growth has been one of the driving forces in recent need not be as harmful to the environment as growth and will continue to play an important some fear; but neither is it always without adverse role in growth and poverty alleviation. consequences. The challenge is to identify the ways in which growth can harm the environment, We are pleased that this report is being issued in so that appropriate actions can be taken to mini- the Envirornment Departrment Working Paper mize damage. Ridwan Ali Chief, Agriculture and Natural Resource Operations South Asia Country Department 1 i Environmnental and Natural Resource Degradation in Intensive Agriculture in Bangladesh ii Contents Foreword i Acknowledgments v Acronyms, Abbreviations, and Glossary vi Executive Summary vii 1. Introduction 1 2. Intensive Agriculture in Bangladesh 3 3. Sustainability of Intensive Agriculture 7 Post-Green Revolution productivity declines 7 Evidence of declining productivity 10 Causes of declining productivity 13 Consequences of declining productivity 17 Farmer responses 18 4. Environmental Problems 21 Pesticide use in Bangladeshi agriculture 21 Main concerns 22 Possibilities for action 24 5. Conclusions 27 References 29 Appendix 31 iii Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh Maps 1. Bangladesh: main rivers and greater districts 2 Figures 1. Total paddy rice production and mean paddy rice yields, 1979-80 to 1993-94 3 2. Intensification in Bangladeshi agriculture, 1978-79 to 1993-94 4 3. Technology changes in the Green Revolution 7 4. Stages in the Green Revolution 8 5. Evidence of yield declines under intensive rice cultivation in long-term fertility experiments at IRRI 9 6. Modem variety rice yields by season, 1983-84 to 1993-94 10 7. Comparison of mean fertilizer application rates on selected modem variety crops, 1979-80 to 1993-94 11 8. Effects of degradation on yields over time 12 9. Relationship between yield and area changes in modem variety boro 12 10. Paddy yield declines in boro crop under different treatments and rotations in long-term yield trials at the BRR1 experiment station, Joydebpur, 1984 to 1994 13 11. Comparison of mean manure application rates on selected modem variety crops, 1979-80 and 1989-90 14 12. Projected changes in total clean rice production and per capita clean rice availability, 1994-95 to 2013-14, under a range of assumptions about annual growth in modem variety yields 17 Tables 1. Change in mean rice area and yields by variety and season, 1979-80 to 1993-94 5 2. Nutrient balances observed in experiment station trials and farmers practices 15 iv Acknowledgments I would like to thank all those who helped in the comments. The report benefitted from extensive preparation of this report, and particularly Ernst consultations with numerous scientists and others Lutz, who accompanied me on the initial mission who have worked on agriculture in Bangladesh in and provided considerable assistance, especially the last decades, in Bangladesh itself, within the on the impact of pesticides on the environment, Bank, at CGIAR centers such as iRRI and cIMMYr, and John McIntire, for overall guidance and and with farmers interviewed during several field support. John Dixon and William Magrath were trips in the fall of 1994 and early 1995. A pre- peer reviewers and provided valuable feedback. liminary draft was discussed at a Workshop held John Baffes, Derek Byerlee, Rashid Faruqee, Hank in Dhaka on March 15, 1995. Gassner, and Keith Pitnan also provided detailed v Acronyms, Abbreviations, and Glossary BADC Bangladesh Agricultural Development Corporation BARI Bangladesh Agricultural Research Institute BBS Bangladesh Bureau of Statistics BPH Brown Plant Hopper BRRI Bangladesh Rice Research Institute CIMMYT International Maize and Wheat Improvement Center DAE Department of Agricultural Extension FAO Food and Agriculture Organization FCD Flood Control and Drainage GDP Gross Domestic Product IFDC International Fertilizer Development Center IPM Integrated Pest Management IRRI International Rice Research Institute NEMAP National Environmental Management Action Plan NGO Non-Governmental Organization NMIDP National Minor Irrigation Development Project TFP Total Factor Productivity UNDP United Nations Development Programme ai Active ingredient (of a pesticide) Aman Rice planted during the wet season and harvested in the dry season Aus Rice planted during the dry season and harvested in the wet season b. Aman Broadcast Aman Boro Rice planted in the dry season t. Aman Transplanted Aman Rabi Dry season ha hectare mt metric tonne ,ug rmicrogram (one rnillionth of a gram) kg kilogram 1 Litre vi Executive Summary Bangladesh has made substantial progress Stagnant or declining yields in the context of towards achieving its goal of foodgrain self- rising inputs indicate that land degradation is sufficiency. This achievement has been based on reducing productivity; if increases in input use a substantial intensification of agriculture: modem had not counteracted the effects of degradation, rice varieties now account for almost half of the yields might have fallen even further. This rice area; an increasing proportion of land is evidence is consistent with patterns of yield double- or triple-cropped; and use of chemical change in other Green Revolution countries, many fertilizers has doubled since the early 1980s. of which have also experienced a slowdown in the Continued and accelerated agricultural growth rate of growth of production and yield. will require intensification to continue. There has been rising concern, however, that intensive Causes of declining productivity. There is consi- agriculture may not be sustainable and that it may derable debate over the exact causes of declining be damaging to the environment or to other productivity. In Bangladesh, the most likely cause productive sectors-such as fisheries-parti- is nutrient imbalances. High-yielding modem cularly through water pollution. This report varieties are far more demanding of soil nutrients examines these concerns. than local varieties had been, a problem worsened by the increasing prevalence of multiple cropping. Sustainability of intensive agriculture Chemical fertilizer use has increased, but not sufficiently to compensate for the higher rates of Evidence of declining productivity. The overall offtake and has been offset by reductions in picture of rising average yields and rising applications of farmyard manure, which is in production in Bangladesh has tended to increasing demand for use as fuel. In general, camouflage evidence of declining productivity. applications of nitrogen are adequate but those of Farmers often claim that yields have been other nutrients often are not. Changes in soil phy- declining and that higher fertilizer applications sical and chemical properties, such as changing are necessary to maintain yields. Analysis of data quality and quantity of organic matter and on yield trends at the district level shows that, formation of a plowpan, also play a role in despite rising input levels, yields have been de- declining yields. Because of its particular charac- clining or stagnant on about two-thirds of the area teristics, on the other hand, Bangladesh has not planted to modem varieties in the boro season in experienced the irrigation-related problems, such the last decade, and stagnant throughout the as declining availability of water, salinization, and country in the araan season. Yield declines are waterlogging, which have been an important strongly associated with the length of time that reason for decreasing productivity in areas such as intensive production practices have been the Pakistani Punjab. There is some evidence of a employed in each district. The results of long- build-up of insects and diseases, but this problem term trials by the Bangladesh Rice Research does not appear to be a major factor in current Institute (BRRI) also indicate that intensive rice yield declines. cultivation can result in declining yields, even under good management and with full recom- Consequences of declining productivity. The switch mended doses of all nutrients being applied. from traditional to modem varieties is essentially Environmental Economics Series VII Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh complete in the dry season and well underway in the farm level. Problems persist, however, in the the wet season. While further expansion of supply of urea, for which a government parastatal irrigation is possible, it will become progressively is the only source. The problems in urea supply more difficult. If current trends continue, the need to be addressed and must be prevented from scope for production growth arising from resultinginback-trackingon liberalization efforts. intensification alone will be exhausted in about the year 2000. Further growth will then depend Environmental problems on improvements in productivity. Since the yield gap between farmer yields and potential yields In addition to sustainability problems, concern has remains high, even in districts which have not been expressed that intensive agriculture harms experienced yield declines (farmer rice yields in the environment more generally. In particular, the dry season, for example, range from about 3 to the rapid increase in the use of pesticides is 4 mt/ha while experiment station yields are thought to (i) adversely affect the health of farm generally in the 6 to 7 mt/ha range), the potential workers and others exposed to pesticides; and (ii) for such growth remains substantial. contaminate ground- and surface water, harming downstream users of that water and damaging Farmer responses. Because yield declines affect inland fisheries. Pesticide use fell in the late them directly, farmers have incentives to respond 1970s, when subsidies were removed, but has to them and are in fact doing so. Farmers are since increased again; sales of pesticides doubled using higher and more balanced fertilizer appli- in the second half of the 1980s. About 70 percent cations. In particular, many have recently begun of pesticides are used on rice. Usage is heaviest using fertilizers supplying micro-nutrients such as on boro, which received over 50 percent of sulphur and zinc. They are also undertaking pesticide applications on rice, by value, in 1989- other activities, such as fuelwood planting, which 90. Nevertheless, the amounts used per unit area will in the long run help to relax some of causes of and the total area affected are both relatively declining productivity. There is considerable small. In 1989-90, only about 10 to 20 percent of scope for the research and extension service to the area planted to modern variety rice was assist farmers in developing appropriate re- treated. Pesticide use on rice is mainly reactive sponses. This will require a re-orientation of rather than prophylactic-applications are only research efforts towards an increased focus on made upon detecting insect infestations in the site-specific conditions, on long-term research, fields. Use of insecticides on vegetables follows a and on cropping systems and practices used by pattern almost diametrically opposed to that farmners. The days of "blanket" messages which found in rice. It is common to spray vegetables are equally applicable to all farmers are over; what such as eggplant and country beans several times farmers need is assistance in fine-tuning their a week. There are indications that resistance to cropping practices to improve yields and avoid pesticides is well established among several degradation, not fundamental qualitative changes. vegetable pests. This effort will also require a much improved extension system, which must not only deliver Health problems. The toxicity of pesticides- information to farmers but also convey infor- threatens the health of users. A study of rice mation back to the research system on farmers' farmers in the Philippines, for example, found needs and constraints. Efforts are already under- health costs associated with pesticide use to be so way to achieve these aims, but much remains to high as to completely outweigh any benefits be done. Farmers will also need reliable and arising from pesticide use. The level and nature of timely access to inputs, including irrigation, pesticide use in Bangladesh, however, differ fertilizers, and credit. The liberalization of significantly from those in the Philippine study irrigation in the late 1980s was largely responsible area, except in the case of vegetable farmers. In for the sharp increase in growth in that period, rice production, pesticide use is widely dispersed and liberalization of the fertilizer distribution and doses are low, so exposure is low. In system during the 1980s has resulted in sub- vegetables, doses are high and applications stantial improvements in fertilizer availability at frequent. Vegetable production is concentrated in viii Environment Department Papers Executive Summary a few areas and among a subset of farmers, so the has already adopted policies which help reduce population exposed is small but at high risk. the risk of substantial environmental problems Vegetables are often grown close to the household arising from pesticide use. Nevertheless, a num- (because of their management intensity) and next ber of relatively simple policies and reforms could to ponds or waterways (to facilitate irrigation) reduce pesticide use even further and make it thus creating the potential for exposure of women safer. Reform of the regulatory framework would and children and for water contamination. help ensure that the pesticides that are used are Consumption patterns also create a very signif- safe both for their users and for the environment. icant potential for pesticide residues onvegetables The current system is slow, excessively concerned affecting the health of consumers. with pesticide effectiveness rather than safety, and consumes valuable research resources. In many Water pollution. Use of agro-chemicals can also instances, it has resulted in older, more toxic result in health problems through pollution of pesticides such as heptachlor (an organo-chlorine) drinking water by residues. The flushing effect of continuing to be used because they are registered, annual floods and monsoon, however, limits the even though safer pesticides are available and in danger that residues will accumulate. Recent tests widespread use worldwide. Increased use of Inte- of groundwater taken from village hand-pumps grated Pest Management (IPM) practices would found traces of pesticides in only ten of 78 allow the low current levels of pesticide use to be samples, despite the samples having been drawn reduced substantially without adverse conse- from areas considered most at risk of contam- quences for agriculture. An FAO pilot project has ination. All the pesticides detected were longer- led to 85 percent reductions in pesticide use lived organo-chlorines, whose use is now banned; among trained farmers and slight yield increases. no traces of the moderately persistent organo- However, the complexity of the knowledge phosphates which account for the bulk of current required for successful use of IPM practices has pesticide use were found. Greater evidence was limited diffusion beyond the trained group. found of nitrate contamination, but even the Given the cost of this extension, the relatively low highest concentrations found were below WHO rate of pesticide use in rice, and the general safe drinking water guidelines. The degradation weakness of the extension service, it may prove of inland capture fisheries has been attributed to beneficial to look for subsets of IPM which could a variety of factors, including overfishing, modifi- be extended more simply, with more complete IPM cations to water flows resulting from water training being concentrated on areas with the control programs, draining of beels for use in highest pesticide use. There is also an urgent need agriculture, and the effects of agricultural and for research on IPM techniques for vegetables. industrial pollution. The relative importance of these factors is impossible to determnine from Conclusions currently-available data, however. Given the low overall level of use, however, pesticides appear Given the scarcity of land and the continued unlikely to have been a major cause of the decline. growth of population, there is no alternative but to continue intensifying agricultural production in Needfor intervention. Except for use on vegetables, Bangladesh. As currently practiced, however, the available evidence does not suggest the intensive agriculture is degrading the soil existence of significant current problems resulting resource base, posing a threat to its sustainability. from pesticide use in agriculture. It is feared that These sustainability problems result primarily the extent of problems might increase as from the difficulties of learning how to manage intensification continues. However, more recent complex new agricultural production systems, not modern rice varieties have been bred for increased from distortions _or market failures inducing pest-resistance, thus reducing the need for inappropriate behavior. They are not, therefore, pesticide. Newer pesticides also tend to have amenable to traditional forms of intervention lower concentrations of active ingredients and to through manipulation of the price structure or by be less persistent. By lifting the subsidies that it rules and regulations on land use. Farmers have paid towards the use of pesticides, Bangladesh amply demonstrated that they have both the Environmental Economics Series ix Environmental and Natural Resource Degradation in intensive Agriculture in Bangladesh incentive and the will to respond to degradation problems. What they need is a supporting infra- structure that will aid them in doing so. There is considerable scope for the research and extension service to assist farmers in developing appropriate responses to degradation problems. This will require a re-orientation of research and extension efforts towards an increased focus on site-specific conditions, on long-term research, and on crop- ping systems and practices used by farmers. Environmentally, intensive agricultural practices appear to pose much less of a threat than is sometimes feared, with the important exception of pesticide use on vegetables. Here too, what problems there are generally do not result from the impact of distorted policies. Several relatively simple interventions can help ensure that they do not grow and may reduce them further. The most pressing need for action is in vegetables, where numerous factors indicate a high risk of damage to the health of farmers and consumers and to the environment more generally. Research is urgently needed to determine the magnitude and exact nature of problems-including threats to the health of farmers and their families, of pollution of waterbodies, and of pesticide residues on food- and to develop appropriate solutions. Reform of pesticide regulation will help to ensure that pesticides in use are safe for users and the envi- ronment. x Environment Department Papers 1. Introduction Substantial increases in irrigated area and use of however, that the actual impact of pesticide use modem rice varieties have led to rapid production has been mninor, since use is both low and growth in Bangladesh in the last decade. Con- dispersed. Furthermore, numerous trends suggest tinued and accelerated agricultural growth, which that environmental problems tied to pesticide use is important both for national economic growth are unlikely to increase substantially; some simple and for poverty alleviation, will require further policy interventions can help ensure that this is intensification. There has been rising concern, the case. Pesticide use on vegetables is an excep- however, that intensive agriculture may be under- tion to this general trend; heavy and frequent mining Bangladesh's natural resource base and its applications of pesticides on vegetables suggest environment. This concern has two components. considerable danger of both health problems and The first is that intensive agriculture may not be water pollution. sustainable. The second is that it may be dama- ging to the environment or to other productive The analysis is based on a wide variety of infor- sectors-such as fisheries-particularly through mation sources, on extensive consultations with water pollution. Various aspects of these concerns numerous scientists and others who have worked feature prominently in the recently-prepared on agriculture in Bangladesh in the last decades, NationalEnvironmentalManagementActionPlan and on field trips to agricultural areas. Compar- (NEMAP). isons to the experience of other countries which have adopted similar practices are often illumi- This report examines these concerns. Section 2 nating, although due account has to be taken of describes the intensification of agriculture which Bangladesh's particular characteristics. In many has taken place in Bangladesh. Intensification has instances, data are insufficient to establish clearly been the driving force behind increases in the magnitude of environmental and natural production and average yields. Section 3 exa- resource problems or to determnine conclusively mines the sustainability of intensive agriculture in the relative importance of various contributing Bangladesh. Numerous sources of evidence show factors. Sufficient data are available, however, to that, despite rising input levels, modern variety arrive at a broad understanding of the relative yields have been declining in many areas. Yield importance of environmental andnatural resource declines have been particularly marked in areas problems and to delineate priorities for action. which have been under intensive cultivation the Throughout the report, the emphasis is on rice longest. Nutrient depletion problems seem to be production and cropping systems in which rice the main cause of declining yields, although other dominates, since rice is by far the most important factors have also contributed. These degradation crop in Bangladeshi agriculture, accounting for problems are to some extent self-correcting, about 50 percent of agricultural output. however, since farmers have incentives to respond to them, and are in fact doing so. Section 4 The scope of this report is limited to natural discusses the environmental consequences of resource and environmental problems resulting intensive agriculture. Rising use of pesticides has from intensive agriculture. Accordingly, it does led to fears of adverse health consequences and of not cover several important natural resource and water contamination. The evidence suggests, environmental problems relevant to the agri- Environmental Economics Series 1 Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh Map 1 Bangladesh: main rivers and greater districts N Dinajpur . .n f ' x . ~~RangptirX ; J0mallpur Bogra f ' Mymensingh i R , Rashahi ':.. .: L--- - R~4shaJuDHAKA .. Kishoreganj~ ,Tangail INDIA DHAKA , - . ~DHAKA, * National Capital !¸ushtia, INDIA .Greater District Boundaries ..Corn d la, Division Boundaries Jessore t Faridpur 1 ¢ Intemational Boundaries flessore Fa lAISO25 50 75 F ''. 8-" ., CH=4G....................... r ? j Barisk4C - NaGON :4 .f es f BARISA ;i | cultural sector, including (i) the possible effects of the extent that it is directly linked to agricultural measures to be implemented under the Flood intensification. Action Plan, (ii) bio-diversity and critical habitat conservation, (iii) erosion and deforestation pro- blems, and (iv) problems tied to shrimp culti- vation. The problem of the decline of inland capture fisheries is likewise not treated, except to 2 Environrnent Department Papers Executive Summary 2. Intensive Agriculture in Bangladesh Agriculture continues to be the most important * Adoption of modem rice varieties (MVs) has sector in Bangladesh's economy, contributing increased substantially, with about half of the about 36 percent of GDP and employing about 60 rice area now planted to modem varieties percent of the labor force. Bangladeshi agriculture (Figure 2, Panel A). Modem variety adoption has made considerable strides in the last decade. has been particularly rapid in boro, where it is Real agricultural GDP has grown by about 2 almost complete. Increased use of modem percent annually, with foodgrains leading the varieties, especially in aman and boro, has been growth. Both production and mean yields of rice the driving force in increasing both total have risen consistently (Figure 1). This growth production (Figure 1, Panel A) and average has resulted in foodgrain self-sufficiency being yields (Panel B). The shift to modem varieties reached in good years-an objective few had has been accompanied by a seasonal redistri- thought achievable. bution of production, with boro growing rapidly at the expense of aus (Table 1). Production increases have resulted from a sub- stantial intensification of agriculture rather than * The irrigated area has expanded rapidly, espe- from increases in cultivated area; indeed, the area cially after the liberalization of minor irrigation planted to rice has fallen slightly in the last in the late 1980s (Figure 2, Panel B). Most irri- decade. Intensification has taken several formns: gated area is devoted to boro production. Figure 1 Total paddy rice production and mean paddy rice yields, 1979-80 to 1993-94 A. Total paddy rice production B. Mean paddy rice yields 30 n - Modern 25s Toal Varieties 3- _20 H0 Mean o15- 2 2- Local 10 - ~~~~~~~~~~~~~~------Varieties 10 - E 5- 0 1979- 1981- 1983- 1985- 1987- 1989- 1991- 1993- 1979- 1981- 1983- 1985- 1987- 1989- 1991- 1993- 80 82 84 86 88 90 92 94 80 82 84 88 88 90 92 94 M MV Aus E MV T. Aman Ei MV Boro Source: BBS Data Environmental Economics Series 3 Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh Cropping intensity has grown substantially, * Use of chemical fertilizers has doubled since the with an increasing proportion of land being early 1980s(Figure2,PanelD),althoughgrowth double- or triple-cropped (Figure 2, Panel C). has slowed in recent years. The progressive This growth in intensity was driven by in- liberalization of fertilizer distribution during the creased cultivation during the dry (rabz) season, 1980s, helped increase fertilizer availability made possible by the growing availability of throughout the country. irrigation. Increased cropping intensity has also led to a concentration of rice production in areas There is still room for further growth along most suited to it and the release of some land Bangladesh's current path of agricultural inten- formerly planted to rice to other crops. sification. Despite the growth experienced in the Figure 2 Intensification in Bangladeshi agriculture, 1978-79 to 1993-94 A. Use of modern varieties, 1979-80 to 1993-94 B. Irrigated area, 1979-80 to 1990-91 i00 - 3.5 - (D ~~~~~~~~~~~~~~~~~3.0- 75 - 2.5 - 2.0 50 -~~~~~~~~~~ CL ~~~~~~~~~~~~~~~~~~1.5 co 25 . 0-0 ~~~~~~~~~~~~~~~~0.5 0 0.0 1979 1981 1983 1985 1987 1989 1991 1993 1978- 1980- 1982- 1984- 1986- 1988- 1990- -80 -82 -84 -86 -88 -90 -92 -9 79 81 83 85 87 89 91 E Boro E Other crops Source: BBS Data Source: BBS Data C. Cropping intensity, 1981-82 to 1991-92 D.Total fertilizer use, 1979-80 to 1993-94 100 -2.5- 80 - - 20 40- 1.0 i111 ° 20- im 0.5~ - l 3 |||| 0 0.0- 1981- 1983- 1985- 1987- 1989- 1979- 1981- 1983- 1985- 1987- 1989- 1991- 1993- 82 84 86 88 90 80 82 84 86 88 90 92 94 Single E Double M Triple-Cropped Urea TSP E MOP = Gypsum Source: BBS Data Source: Fertilizer Association 4 Environment Department Papers Executive Summary Table 1 Change in mean rice area and yields by variety and season, 1979-80 to 1993-94 Area ('000 ha) Mean Yields (mtlha) 1979 1991 Change (%) 1979 1991 Change (%) Variety Season -82 -94 Total Annual_ -82 -94 Total Annual Local Aus 2,645 1,371 -48 -5.5 1.17 1.35 16 1.3 b. Aman 1,584 886 -44 -4.8 1.36 1.44 6 0.5 t. Aman 3,493 2,644 -24 -2.3 1.71 1.98 16 1.2 Boro 414 265 -36 -3.7 2.07 2.05 -1 -0.1 Total 8,136 5,166 -37 -3.8 1.48 1.73 16 1.3 ModemAus 480 394 -18 -1.6 2.91 2.65 -9 -0.8 t. Aman 483 1,895 293 11.4 2.99 3.22 8 0.6 Boro 706 2,297 226 9.8 3.88 3.88 0 0.0 Total 1,668 4,585 175 8.4 3.34 3.50 5 0.4 Total Aus 3,125 1,767 -43 -4.7 1.43 1.64 15 1.2 b. Aman 1,584 886 -44 4.8 1.36 1.44 6 0.5 t. Aman 4,426 4,862 10 0.8 1.93 2.51 30 2.2 Boro 1,204 2,605 116 6.4 3.20 3.69 15 1.2 Total 10.339 10.120 -2 -0.2 1.84 2.57 39 2.8 Note: Pajam area is included in the total but not shown separately Source: Computed from BBS data past decade, only about half the land classified as suitable or very suitable for irrigation in the rabi season is irrigated. Intensification, however, cannot continue inde- finitely. The switch from traditional to modem varieties is essentially complete in boro and well underway in aman. Further expansion of irri- gation is possible, but will become progressively more difficult. Modem varieties and irrigation were initially adopted on land most suitable to them; further increases in their area are likely to be costlier and may not have as large a yield effect as previous ones. As in other Green Revolution countries, therefore, growth in production and average yields will require increasing the yields of modem varieties. Unfortunately, there is increasing evidence that intensive agricultural practices may be degrading the natural resource base on which agricultural production depends. Yields of modem varieties, far from increasing, may actually be declining despite higher input levels. Environmental Economics Series 5 Envirorunental and Natural Resource Degradation in Intensive Agriculture in Bangladesh 6 Environmnent Departnent Papers 3. Sustainability of Intensive Agriculture This section discusses the sustainability of the Post-Green Revolution productivity declines intensive agricultural production practices being increasingly used in Bangladesh. First, the expe- Figure3providesaframeworkforexamininghow rience of other Green Revolution countries is the Green Revolution affects the technological reviewed. A widely noted slowdown of rates of possibilities available to farmers. agricultural growth throughout these countries is raising concern over the sustainability of intensive * Traditional local varieties use low input levels agricultural practices; their experience provides and have low returns. This technology is valuable insight into the problems being expe- illustrated by curve LV. Optimal fertilizer use, rienced by Bangladesh. Second, available evi- given the relative price of outputs and inputs, is dence on declining productivity of agriculture in FT, which produces output YT3 Bangladesh is reviewed. Farmer reports, national statistics, and the results of long-term cropping * Introduction of modern varieties increases the trials at BRRI all indicate that yields are stagnating yield potential dramatically, as ilustrated by and productivity is declining,' although data are curve MVp, but requires higher input levels. insufficient to establish precisely the rate of Under this technology, optimal fertilizer use is decline. The causes and consequences of this F*, which produces output Yp*. decline are then discussed, as are the farmers' responses and ways in which government policies can help. Figure 3 Technology changes in the Green Revolution Yield YPMAX --- v Technical Inefficiency 1 Allocative to* ------ -- MV0 Inefficiency L/ LV Yr ______________________ ___________ ______ Fertilizer FT F0 * F* FPM,~ Application ET nO FOa EPi SPMAX Environmental Economics Series 7 Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh * However, because of inexperience with the new and then from increasing technical efficiency systems, in practice the achievable yield from [Byerlee, 1992]; these are shown as movements modem varieties is initially lower, as illustrated along a production function and as upward shifts by curve MVo. Under this technology, optimal of the achievable production function, respec- fertilizer use is Fo*, which produces output YO*. tively. As illustrated in the figure, the stages of improving allocative and technical efficiency over- * Farmers often use inputs below their optimum lap, with imnprovements in allocative efficiency level because of limited availability, risk dominating initially and improvements in tech- aversion, or problems such as credit constraints, nical efficiency becoming most important later. which prevent them from financing high input levels. So farmers use inputs at a level such as There has been a widely-noted slowdown in the Fe. which produces output YO. rate of growth of production and yield in Green Revolution countries [Bouis, 1993; Byerlee, 1992; Using this formulation, the 'yield gap' between Pingali, Moya, and Velasco, 19901. This slow- achieved and potential yields can be decomposed down is partly due to the completion of the into two components. The first gap results from transition from traditional to modem varieties. shortfalls in allocative efficiency (Yo* - Yq). This gap The potential for further expansion of modem is between achieved yields and optimal yields varieties and irrigation is limited in most given the current technology. The second gap countries. The slowdown is also partly due to the results from shortfalls in technical efficiency reduction in rice prices which has resulted from (YP* - Y0*). This gap is between optimal yields the earlier production increases. At the same under the current technology and potential time, governments have often retreated from the yields.3 policies of high input subsidies that many adopted in the early stages of the Green Green Revolution yield increases have generally Revolution. These adverse price changes have occcured in a series of stages, as illustrated in reduced incentives to increase yields. Figure 4. Adoption of modem varieties led to a significant immediate yield increase. Following There is growing evidence, however, that more this transition, further yield increases have tended fundamental problems are also at work. The rate to come first from increasing allocative efficiency of growth of total factor productivity (TFP) in Figure 4 Stages in the Green Revolution Yield Yield MVP Increase in --------------------- MV2--------------------------------------------- --Efficiency: Y2 1------------ ; / = MV- Technical Allocative YO - ----- /Ht---- ------------------------------------- ----------------- Adoption I of Modem Varieties LV Y T - ---- ---- ---- ---- ---------------- - ' _____________________ _____________ Fertilizer |_Tim e FT FO FX F2 Application 8 Environment Department Papers Sustainability of Intensive Agriculture India fell from 1.22 percent annually during 1965 nutrient deficiencies, therefore, are a common to 1975 to 0.98 percent annually in the following problem. Intensive agricultural practices can decade; in Pakistan, the decline was even sharper, also result in changes in soil physical and with an annual TFP growth rate of 1.86 percent in chemical properties. In irrigated rice, the for- the period 1965-75 giving way to a 0.36 percent mnation of a plowpan is a frequent problem, annual decline in 1975-85 [Rosegrant and Evenson, impeding root growth and limiting access to 1992]. A decomposition analysis of sources of nutrients. Irrigation can also cause salinization growth in the Pakistani Punjab shows that growth or waterlogging. The yield declines observed in was about 20 percent lower than would have been the IRRI long-term cropping trials are thought to expected from the continuing switch from result from anaerobic conditions in irrigated rice traditional to modem varieties, the growth in production leading to changes in soil physical input levels, and the genetic improvements in the and chernical properties which make it less able newer modem varieties which have become to supply nutrients to growing crops [Cassman available [Byerlee and Siddiq, 1994]. These statis- et al., forthcoming]. tical data are reinforced by evidence from long- term cropping trials conducted at IRRI and * Degradation of ancillary resources. Most Green elsewhere in Asia, which found persistent Revolution success stories depend heavily on declines in yields over time even under optimal irrigation. The water used for irrigation is itself mana-gement practices [Cassman et al., a natural resource, which can be degraded forthcoming]. Figure 5 shows the overall yield through misuse or overuse. In many areas, trends observed in the long-term IRRI trials and extraction of groundwater is greater than the underlying shift in the fertilizer response replenishment, leading to a drawdown of the function. aquifers which threatens the long-term sustain- ability of agriculture and increases pumping The literature on the Green Revolution suggests costs. Over-use of water has often been en- several possible explanations for productivity couraged by substantial subsidies to irrigation, declines in intensive rice cultivation systemns: either directly or indirectly by subsidies to the energy used for pumping. In some areas, * Degradation ofthe soil resource base. The intensive degradation of irrigation infrastructure cropping systems introduced by the Green resulting from insufficient maintenance has also Revolution impose much heavier demands on begun to affect production [Rosegrant and crop nutrients than traditional systems had; Pingali, 1994]. Figure 5 Evidence of yield declines under intensive rice cultivation in long-term fertility experiments at IRRI A. Changes in fertilizer response function B. Trends in yield of highest-yielding varieties Yield Yield (mt/ha) (mt/ha) 10 ~~~~~~~1970-72 8 872 8 ~ Dry 6 11989-91 Season 4 ~~~~~~~~~~~~~~~~~~Wet 2N Input 2 Season o 50 100 150 200 (kg/ha) 1 965 1970 1975 1980 1985 1990 Ya Source: Cassman et al., 1995 Environmnental Economics Series 9 Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh Build-up of pest pressure. The intensive pro- perceptions of declining productivity have also duction practices introduced by the Green been noted in numerous surveys conducted in Revolution provide an ideal environment for recent years. A pilot survey of Kaliganj Thana in pest build-up. Multiple cropping and over- Gazipur District, for example, found that 69 lapping seasons facilitate the perpetuation of percent of respondents in one village and 17 pest populations, and the existence of large percent of respondents in another thought fertility homogeneous area favors the build-up of spe- was declining; only 5 percent in either village cialized pests. Moreover, early modern varieties thought it was increasing [SFFP, 1994]. Alauddin were often highly susceptible to local pests. and Tisdell [1991] report that all farmers in the Estimates of production loss to pests in Asia village of South Rampur, in Comilla District, and range as high as 20-30 percent annually. In 88 percent of farmers in the village of Ekdala, in response to these problems, regular prophy- Rajshahi District, said they had to apply more lactic pesticide applications were a standard fertilizer than before to maintain yield. Saunders part of the early Green Revolution package and [1990, 19911 found that up to a fifth of wheat pesticide sales were often subsidized. The farmers surveyed in the Jessore/Kushtia area and consensus of opinion today is that this approach in Dinajpur thought rice and wheat yields were was mnisguided on two counts.4 First, large- declining.5 scale and frequent use of pesticides bred resistance into target pests, thus requiring ever- Statistical data. Since 1972-73, the Bangladesh increasing doses to be used. Second, pesticides Bureau of Statistics (BBS) has collected data on often proved equally lethal to beneficial pests production of rice and other important crops from which preyed on crop pests. In many cases, a sample of 5,000 five-acre plots nationwide. therefore, pesticide use actually resulted in These data allow a close analysis of modem lower yield as crop pests, freed of their natural variety yields.6 Figure 6 shows average modem predators, multiplied without constraints. variety rice yields from 1979-80 to 1993-94 in each of the rice production seasons.7 The changes in Evidence of declining productivity yields in this period are also summarized in Table 1. Although average yields increased by almost 40 The overall picture of rising average yields and percent during this period (equivalent to an rising production in Bangladesh has tended to average annual increase of about 2.8 percent), camouflage evidence of stagnating and declining modemrvariety yields have experienced very little productivity. Despite occasional expressions of growth, if any. The changes in modem variety concern from scientists and policy-makers (eg yields shown in Table 1 are not significantly dif- Bhuiyan [1991]), the problem has received ferent from zero in any of the seasons.8 relatively little attention. Indeed, the possibility that Bangladesh's soil resources are being de- graded received surprisingly little attention in the Figure 6 recently-prepared NEMAP. Yet evidence from Modem variety rice yields by season, farmer perceptions, national statistics, and long- 1983-84 to 1993-94 term cropping trials at the Bangladesh Rice Research Institute (BRRI) all indicate that, in many 5 parts of the country, productivity is stagnating or declining. 4 j Boro 3 - ------- -Aman Farmer perceptions. There is a widespread percep- 3 Au s......-- -Aua tion of declining productivity among farmers. All 2u 2 - farmers interviewed during field trips to the E northwestern and southwestern regions of Ban- gladesh in the fall of 1994 claimed that yields had o been declining and that higher fertilizer appli- 1979 1981 1983 1985 1987 1989 1991 1993 cations were necessary to maintain yields. Farmer -80 -82 -84 -86 -88 -90 -92 -94 Source: BBS data 10 Environment Department Papers Sustainability of Intensive Agriculture Use of district-level data allows yield trends to be increased markedly in the last decade. As in other examined in more detail. Regression analysis of Green Revolution countries, input use was the time trends of modern variety yields at the initially very low. Cash-constrained, risk-averse district level shows considerable differences Bangladeshi farmers had difficulty in financing between yield trends in different districts (see the required high input levels. Moreover, in the Appendix). Eleven districts, representing just early stages of the Green Revolution in under half the boro area, have experienced Bangladesh the fertilizer distribution system was declining yields in the last decade, while yields controlled by a state monopoly, the Bangladesh have stagnated in a further three districts. In Agricultural Development Corporation (BADc), several instances, increasing yield trends at the whose performance proved woefully inadequate. beginning of the period became declining trends Although there is evidence that some allocative in more recent years. Only seven districts, inefficiency persists, input use has gradually representing just over a third of the area currently increased over time. Despite these increases in planted to modem variety boro, have experienced fertilizer use, boro yields have stagnated or rising yields in the last decade, and in two of these declined in two-thirds of the area planted to them, the rate of yield increase has slowed in the last and aman yields have stagnated throughout five years. In aman, on the other hand, modem Bangladesh. variety yields appear to have stagnated in most districts; few yield trends are significantly That yields should stagnate or decline despite different from zero. rising input use indicates that productivity is falling and strongly suggests that land degra- The implications of yield stagnation or decline for dation is reducing achievable yields (Figure 8). If productivity are severe, since these trends have increases in input use had not counteracted the occurred despite rapid growth in the use of effects of degradation, yields might have fallen chemical fertilizers. Figure 7 shows that use of even further. Examination of district-level data chemical fertilizers on modem variety crops has provides further evidence of degradation. * District-level patterns of yield change are Figure 7 negatively associated with the length of time Comparison of mean fertilizer application that intensive production practices have been rates on selected modem variety crops, employed (Figure 9, Panel A). Since increases 1979-80 to 1993-94 in technical efficiency through learning-by- doing and increases in allocative efficiency 400 should both be leading to gradual yield 1 increases in areas where modem varieties are well established, the opposite might have been 300- expected. That yields should be either stag- nating or declining suggests that degradation is sm 200- lti 11 offsetting any positive effects of increasing 0) i~~~~~~~~ ~ ~ ~ ~ ~technical and allocative efficiency. 100 i* * District-level patterns of yield change are positively associated with rapid growth in the area under boro (Figure 9, Panel B). Again, the 0- opposite might have been expected: since 1979 1989 1979 1989 1979 1989 modem varieties are likely to be planted first in -80 -90 -80 -90 -80 -90 areas most suitable to them, the average suit- MV t. Aman MV Boro Irrigated ability of land under modem varieties tends to Rice Rice MV Wheat decline as the area expands. Other things equal, Urea 10 TSP MOP this should lead to lower average yields. That Sources: 1979-80: IFDC; 1989-90: BBS average yields should increase under these Environrmental Economnics Series 11 Environmnental and Natural Resource Degradation in Intensive Agriculture in Bangladesh Figure 8 Effects of degradation on yields over time Yield Yield MVp * Increasing Inputs o Constant Inputs MV0 MV1 LV Fertilizer FT FO F1 F2 Application Time conditions suggests that the land being brought other Green Revolution countries: yields initially under modern varieties is less degraded than rise after modem varieties are adopted, reflecting land which has been cultivated intensively for a improvements in allocative efficiency and the longer period. benefits of learning-by-doing. After intensive practices have been in use for some time, how- These results suggest that intensive agriculture in ever, degradation begins to set in and yield Bangladesh is following a path similar to that of growth is slowed or reversed. Figure 9 Relationship between yield and area changes in modern variety boro A. Relationship between yield change and B. Relationship between yield change and length of time under intensive practices growth in area under intensive practices ,D_120 25 r_> 20l tgEoX t 0) a2 -1.S >) CCU 6 1l9 Co CO o 0~~~~~ (1.73) (-2.41) (-2.68) (3.20) 1Lu2 Eniomn Deprten Paer 2 a-) -1.5-1 -0.5 00.51 1.5 -2 -1 0 1 2 IL Ca Average annual yield change Average annual yield change 1983-84 to 1993-94 ()1979-80 to 1993-94(% Adj. R2 = 0.203 Adj. R2 = 0.316 Yield Change = 0.760 -00181 Initial Area Yield Change = -1. 19 + 0.118 Area Change (1.73) (-2.41) (-2.68) (3.20) Source: Authors' calculations 12 Environmnent Department Papers Sustainability of Intensive Agriculture Experimental station data. Data from long-term Causes of declining productivity yield trials at the BRRI research station at Joy- debpur also indicate that intensive rice cultivation As in other Green Revolution countries, there is can result in declining yields (Figure 10), even considerable debate over the exact causes of when full recommended doses of all nutrients are declining yields. This section reviews the avail- applied. In this respect, the BRRI results parallel able evidence on causes of declining productivity those obtained in the long-term cropping trials at in intensive agriculture in Bangladesh. Consi- IRRI (Figure 5 above). The overall yield level is derable evidence points to imbalances in nutrient lower and the rate of yield decline faster when availability as the main source of most of the lower doses of fertilizers are applied. T. aman current productivity decline, but data are paddy yields, on the other hand, do not seem to insufficient to clearly establish the importance of show any clear downward trends in the long- the numerous factors at work. term Joydebpur trials. Nutrient imbalance. High-yielding modem va- These various sources of data demonstrate that rieties are far more demanding of soil nutrients productivity under intensive agricultural than local varieties had been. Demands on practices, using modem varieties and high input nutrients are all the more severe given the levels, is declining in many parts of Bangladesh. increasing prevalence of multiple cropping. Unfortunately, time series of inputs that distin- Although additions of chemnical fertilizers have guish between input use on different crops, on increased substantially (Figures 2D and 7), they different varieties, and in different seasons are not have not increased sufficiently to offset the higher available.? Without such data, the rate and magni- rates of offtake. Moreover, the increase in chemi- tude of productivity decline cannot be measured cal fertilizer applications has been to some extent quantitatively. offset by reductions in applications of farmyard manure, which is in increasing demand for use as fuel (Figure 11). Nutrient balance analysis shows Figure 10 Paddy yield declines in boro crop under different treatments and rotations in long-term yield trials at the BRRI experiment station, Joydebpur, 1984 to 1994 A. Boro-Fallow-tAman B. Boro-Aus-t.Aman 6 - 6 - 198- 196 98 90 92 93- 18- 18-18- 19 NPKSZn - __ NoS No -A NPKS E 3 - N only E applied3 *No - No f 2 2 ------. None I~ ~ ~ ~ ~~~~~~~~~~~~~~~~ ~ore BRR data 1984- 1986- 1988- 1990- 1992- 1983- 1985: 1987- 1989- 1991- 86 88 90 92 94 85 87 89 91 93 * NPKSZn - Full application of all nutrients u NPKS - Full application of all nutrients * No S - All nutrients applied except S * No S - All nutrients applied except S * N only - Only N applied * None - No fertilizer applied Source: BRRI data Envirorunmental Economics Series 13 Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh Figure 11 particular have an adverse effect on yields Comparison of mean manure application (Figure 10). Sulphur and Zinc deficiencies are a rates on selected modem variety crops, common problem in intensive rice production, 1979-80 and 1989-90 because conditions in flooded paddies have a strong negative effect on their availability 2500- [DeDatta, 1981; Tandon, 19901. 2000- Evidence of high fertilizer responses also supports the hypothesis of inadequate fertilization [Islam, 1500- yg8 1990]. If the hypothesis is correct, then it is possible that productivity decline will bottom out at some point even without greater fertilizer 1000- , : >gc m .. applications, since lower yields will mean lower nutrient removal. Evidence from long-term trials 500- in India and some evidence fromBangladesh itself suggests that, at least in the short to medium term, 0- yields can be sustained with appropriate balanced 1979 1989 1979 1989 1979 1989 fertilizer applications. However, the data in Table -80 -90 -80 -90 -80 -90 2 indicate that nutrient applications can be MV t. Aman MV Boro Irrigated unbalanced even when ful recommended doses Rice Rice MV Wheatunaacdeewhnflrcom dddss of each nutrient are applied to every crop in the Sources: 1979-80: IFDC; 1989-90: BBS rotation. Changes in soil physical and chemical properties. that nutrient removals often exceed additions Numerous analysts have pointed to the low (Table 2). organic matter content of Bangladeshi soils and to their deficiency of certain nutrients. But all the a Macronutrients. Applications of nitrogen (N) are data available consist of single observations. generally adequate, reflecting the ready avail- Whether soil physical and chemical properties are ability of nitrogenous fertilizer (urea is pro- in fact declining, and if so at what rate, cannot be duced in the country), its relatively low price, determined. and the easily-observed yield effects of N appli- cations. Applications of phosphorus (P) are also * In general, soil organic matter tends to increase generally adequate. Potassium (K) balance, under Green Revolution-style intensive however, is often inadequate. In the past, production practices. Given the drastic reduc- potassiumhadnotbeenthoughttobeaproblem tions in manure applications and the commnon because many of the sediments from which practice of removing practically all crop straw Bangladeshi soils are derived are potassium- and stubble from the fields (for use as fuel or rich. But intensive removal through high-yield fodder), however, this may not be occuring in rice production is making potassium limiting. Bangladesh. Even if it were, research at IRRI The problem would be less severe if rice straw suggests that changes in the nature of organic were returned to the soil, since most of the matter under long-term irrigated conditions potassium removed by rice is in the straw. make it less capable of supplying nutrients to Straw is in high demand for use as fuel and growing crops; this is thought to be an animal fodder, however, and very little of it is important reason for the continuous decline in returned to the soil; in fact, in many areas even rice yields seen in the long-term trials at IRRI. the crop stubble is removed. Periodic drying of irrigated rice fields appears to play an important role in preventing, or at * Micronutrients. In the case of rice, sulphur (S) least slowing, the emergence of these problems. and zinc (Zn) are the main sources of concern. Since triple-cropping of rice is rare in Bangla- BRRI trials show that sulphur deficiencies in desh, soils are not flooded as long as in some 14 Environment Department Papers Sustainability of Intensive Agriculture Table 2 Nutrient balances observed in experiment station trials and farmers practices Nitrogen Phosphorus Potassium Rotation. Treatment Added Removed Balance Added Remoed Balance Added Removed Balance A. BARI experiment station trials Janokinathpur FRS, Rangpur Wheat - Mungbean - MV t. Aman NPKSZn - NPK - NPK 180.0 192.1 -12.1 160.0 78.1 81.9 100.0 218.1 -118.1 NPKSZn- NlhP%K - N½P½K 180.0 183.4 -3.4 110.0 74.6 35.4 70.0 207.9 -137.9 NPKSZn - N - N 180.0 180.4 -0.4 60.0 74.0 -14.0 40.0 205.6 -165.6 N - N - N 180.0 114.1 65.9 0.0 50.0 -50.0 0.0 143.2 -143.2 Narhatta FRS, Bogra MV Boro - Fallow - MV t. Aman NPKSZN - - NPK 180.0 168.5 11.5 120.0 89.2 30.8 80.0 267.6 -187.6 NPKSZN - - N1A½hK 180.0 147.6 32.4 90.0 78.1 11.9 60.0 234.4 -174.4 NPKSZN - - N 180.0 140.1 39.9 60.0 74.2 -14.2 40.0 222.4 -182.4 N - - N 180.0 122.6 57.4 0.0 64.9 -64.9 0.0 194.7 -194.7 B. Farmer practices Jessore/Kushtia Area Wheat - Aus 147 120 27 95 55 40 45 148 -103 Wheat - Aus - Aman 242 195 47 158 93 65 82 264 -182 Wheat - Aman 198 160 38 124 75 49 64 206 -142 Meghna, Chandpur Irrigation Project Potato - MV boro - MV t aman 250 258 -8 61 44 17 148 339 -191 Potato - MV boro - MV t aman 319 257 62 103 44 59 164 338 -174 Dakatia, Chandpur Irrigation Project LV b. Aman - MV Boro 192 117 75 22 19 3 28 158 -130 WV Boro 151 93 58 32 15 17 22 122 -100 NV t. Aman - MV Boro 175 142 33 32 24 8 22 192 -170 LV t. Aman - MV Boro 128 150 -22 37 25 12 50 204 -154 Sources: BARI from Abedin and Mukhopadhay, 1990; Jessore/Kushtia from Saunders, 1991; Chandpur from ISPAN, 1993 other Green Revolution countries. Under typi- encourages the formation of a plowpan), in the cal current practices with double-cropped rice, long term it can be harmful by reducing the however, soils are dry for only about one month crops' access to nutrients. Periodicallybreaking each year, just before the boro season, which the plowpan by planting deep-rooted crops may not be sufficient. such as jute seems to have a beneficial effect on yields. There is some evidence, for example, The development of a plowpan under irrigated that rice-wheat rotations which include break rice production is often pointed to as an un- crops tend to have fewer yield decline problems favorable change in physical structure. [Saunders, 19911. Although the formation of a plowpan is desi- rable in the short term, since it allows easier Decline in properties of modern varieties. This control of irrigation water (indeed, puddling of hypothesishas two variations. One focuses on the the soil prior to transplanting rice deliberately productivity of the seed. Most farmers plant seed Environmental Economics Series 15 Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh retained from their harvest rather than renewing to outbreaks of insects and diseases. Conditions their seed stock by buying certified seed. No trials are particularly favorable in the southern part of have been conducted on the optimal time for seed the country, where winter temperatures are high. renewal. Questions have also been raised about Infestations by rice hispa have occurred regularly the adequacy of current seed multiplication since 1978 and are thought by some to be practices. The other variation is that the modem increasing. Recent years have also seen outbreaks varieties themselves are losing their vigor. Very of other pests. An outbreak of brown plant hop- little research has been undertaken on this latter per (BPH) in the Rajshahi-Dinajpur area in 1993, possibility, and none in Bangladesh. for example, caused considerable damage. Since pest outbreaks tend to be very irregular, however, Irrigation-related problems. Problems related to it is difficult to determine whether their incidence irrigation, including decreasing availability of and severity are increasing. The relatively low water, salinization, and waterlogging, have been use of pesticides on rice suggests that growing an important reason for declining productivity in resistance is unlikely to be as important a problem some areas, such as the Pakistani Punjab. These as it has become in other countries, such as problems do not appear to be significant in Indonesia. Bangladesh. Unlike some other Green Revolution countries, groundwater in Bangladesh is replen- The various types of problems discussed above ished at rates that exceed current rates of use. The can be categorized broadly as medium-term and overall availability of water is unlikely to become long-term causes of yield decline. a constraint to agricultural production in Bangla- desh. Seasonal problems may develop in certain * Medium-term problems. Problems such as areas, however, if water extraction drives the nutrient imbalances are medium-term problems water table below the reach of some wells at the since their effects, while not immediate, are felt end of dry season. This appears to be happening relatively rapidly. Their effects can also be in parts of the Barind Tract, for example. In times reversed relatively rapidly, once appropriate of drought, the area affected is likely to be larger, changes are made in management practices. as was in fact the case during the 1994-95 boro season. These are temporary problems, however, * Long-term problems. Changes in soil physical which do not cumulate over time. Salinization, and chemical properties, on the other hand, another problem often encountered in irrigated often tend to be longer-term, both in terms of areas, also does not appear to be a problem in the time it takes for their effects to be felt and Bangladesh, except for a small area west of the time it takes to reverse them. The dete- Comilla, thanks to the annual flushing effect of the rioration in soil properties noted in long-term monsoon and of flooding. Waterlogging is experiments at IRRI appears to fall in this latter reported in some areas, but appears to be a con- category. Since the adoption of modem va- sequence of inappropriate construction of roads rieties is more recent and still less complete in and other obstacles to proper drainage, rather Bangladesh compared to most other Green Re- than from misuse of water at the field level. That volution countries, these long-term degrading irrigation-related problems are generally minor effects of intensive production practices are does not mean, however, that water management unlikely to already be having a significant effect, is unimportant. Although the potential exists for except perhaps in areas where modem varieties irrigation to continue to grow, inadequate water were adopted relatively early. These are pro- management could reduce its benefits.10 The blems that Bangladesh may well encounter in situation in coastal areas is particularly delicate, future years, however. since changes to the local hydrology can result in saline water intrusions into both ground and In addition to these medium-term and long-term surface water. problems, there are many short-term problems which affect current productivity without Build-up of insects and diseases. Continuous rice necessarilycausingresourcedegradation. Impro- monoculture creates conditions that are favorable vements in the timing and mode of application of 16 Environment Departmnent Papers Sustainability of Intensive Agriculture fertilizers, in water management, and in culti- change. Output in each period depends on the vation practices would all help improve current total area planted to rice in each season, on the productivity, as would improvements in seed proportion planted to modem varieties, and on quality. The yield gap between farmer yields and the yield of modem and local varieties. The area potential yields remains large, even in districts planted in each season and the proportion planted which have not experienced yield declines. In to modem varieties are assumed to continue other Green Revolution countries, the yield gap changing at the same rate as was experienced in has become very small. In the Ludhiana District the period 1983-84 to 1993-94. However, there are of the Indian Punjab, for example, the gap season-specific limits on the proportion of land between farmer yields and the best yields which can be planted to modem varieties. In the obtained on experiment stations has fallen from boro season, the limit is 100 percent. In the other almost3mt/hato about 1.5 mt/ha [Byerlee, 1992]. seasons, however, the possibility of flooding In the Philippines, the top third of irrigated farns prevents the entire area from being planted to now have yields that are very close to experiment modem varieties. A limit of 80 percent adoption station yields [Pingali, Moya, and Velasco, 1990]. is assumed for both aman and aus. The initial In Bangladesh, in contrast, the yield gap has levels of areas and yields are set at their average remained essentially unchanged. Farmer boro over the three-year period from 1991-92 to 1993- yields, for example, range from about 3 to 4 94. mt/ha, depending on location, while experiment station boro yields are generally in the 6 to 7 The results of the model's projections of total mt/ha range. production over the next 20 years, under a range of assumptions for modem variety yield changes, Consequences of declining productivity are shown in Panel A of Figure 12. If modem variety yields are indeed stagnating, then the Projections of growth under different scenarios of prospects for future production are fairly grim. productivity change indicate the likely Some growth will continue as a result of the consequences of declining productivity for agri- continued switch from local to modem varieties cultural growth. and from changes in the relative importance of the different seasons (in particular, the decline of aus A seasonally and varietally disaggregated model and the growth of boro). But the transition to mo- of rice production is used to project growth dem varieties is already essentially complete in patterns under different scenarios of productivity boro and will be completed in about the year 2000 Figure 12 Projected changes in total clean rice production and per capita clean rice availability, 1994-95 to 2013-14, under a range of assumptions about annual growth in modem variety yields A.Total production B. Per capita availability 251- +2% +1% t80 - +2% 20 - 0% 160 - - +11% -1% -j 0% E 15 -2% 120 -11% .2 80o -2% 60- 5 ~~~~~~~~~~~~~~~40- 20 1993-94 1998-99 2003-04 2008-09 2013-14 1993-94 1998-99 2003-04 2008-09 2013-14 Source: Author's calculations Environmental Economics Series 17 Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh in aman, at current growth rates. Once the tran- Farmers generally have strong incentives to sition in these two important seasons is com- respond to soil degradation problems: a decline in pleted, production growthwill slow dramatically, land productivity affects them directly. And in as can be seen by the kink in the time paths of fact, available evidence suggests that farmers are production. If modem variety yields are stagnant, already responding to degradation problems, by the only source of growth remaining at this point using higher and more balanced fertilizer appli- will be that resulting from the continued relative cations. Inparticular,manyfarmershaverecently decline in aus area and growth in boro areas." If adopted fertilizers supplying micro-nutrients such yields are in fact declining, this effect alone will be as gypsum (a source of sulphur), zinc sulphate, insufficient to counteract the effects of yield and borax (a source of boron). Practically all decline and production will fall. farmers interviewed during field trips to the northwestem and southwestem regions of The implications of these projections for per capita Bangladesh in the fall of 1994, for example, had rice production, assuming that population will begun using gypsum within the last five years. continue to grow at the current rate of about 2 Other surveys have made similar findings (see percent per annum, are drawn out in Panel B of also Figure 2C). Saunders [1991], for example, Figure 12. Unfortunately, if yields are in fact found that over 60 percent of farmers surveyed in stagnating, the current situation of rice self- Jessore and Kushtia districts had begun applying sufficiency will prove very short-lived; by about gypsum to their aman crop, almost all within the the year 2000, per capita production will begin to preceding four years. Farm households are also fall. If yields are in fact declining, this will occur planting large numbers of trees, which will help even sooner. Continued efforts to slow popu- ease the pressure to use cowdung and crop lation growth will help ease this pressure, but are stubble for fuel. Renters and sharecroppers are a unlikely to eliminate it entirely. possible exception to the rule that farmers have strong incentives to respond to degradation The results shown in Figure 12 should be cause for problems; whether they do or not depends on the concern but not for despair. First, they show that specific features of their rental and sharecropping except under fairly pessimistic assumptions of contracts. Evidence from several case studies, current yield changes, continuing adoption of however, indicates that factors such as fertilizer modern varieties and changes in the relative use and modern variety adoption do not differ importance of the different seasons will probably between owner-operated and crop-shared or be sufficient to sustain per capita production at rented land [Taslim and Ahmed, 1994). roughly current levels for several years. There is, therefore, some time for action to be taken. For fanners to respond appropriately to degra- Second, they show that if reasonable rates of yield dation problems, they require infornation on the growth can be achieved, per capita production can long-term consequences for productivity of be sustained and perhaps improved. In light of different agricultural practices and on appropriate the large yield gap that continues to exist between corrective measures. In general, farmers usually current and potential yields, there is still have very good knowledge of their soil conditions considerable room for growth in yields by and of the effects of agricultural practices upon improvements in allocative and technical them. But this is knowledge that results from long efficiency. experience. The intensive agricultural practices being used today are too recent for such Farmer responses experience to have accumulated to any substantial degree. In a dynamic setting such as that intro- Whether the adverse consequences described duced in the Green Revolution farmners depend above come to pass will depend primarily on the much more on information provided by research farmers' responses to degradationproblems. Ulti- and extension systems. Unfortunately, in Bangla- mately, it is farmers who make land use decisions, desh that system does not appear to have, except and not government analysts and planners in the in the broadest qualitative sense, the required capital. information. Greater use of chemical fertilizers is 18 Environment Department Papers Sustainability of Intensive Agriculture part of the answer, but since use has already been nutrient needs rather than standardized guide- increasing, further increases in input use alone are lines. Recent years have seen some moves in this unlikely to be sufficient; better use of existing direction, but much more needs to be done. There inputs is required. This task is more knowledge- have been recent efforts in several countries to intensive and less amenable to simple govermment establish soil testing and advisory services for interventions like the input subsidies of yester- farmers [Hanson, 1994]; the possibility of year. developing such services for Bangladeshi farmers should be investigated. * Research. Research is required into fertilizer needs for long-term sustainable production. As Farmers also need to have the means to respond. discussed above, evidence from cropping trials Unless farmers have reliable and timely access to undertaken by the Bangladesh Agricultural inputs, including irrigation, fertilizers, and credit, Research Institute (BARI) suggests that current they will be unable to increase the efficiency of recommendations are inadequate to maintain their cropping practices. The liberalization of long-term fertility in many cases (as well as irrigation and fertilizer distribution in the late being inadequate in other ways, having been 1980s was largely responsible for the sharp developed primnarily on a single-crop, single- increase in growth in that period. season basis). Research also needs to go beyond documenting yield declines to understanding * Credit. Increased input use will create higher the causes of these declines. Perhaps as impor- financing requirements. An efficient and effec- tant as fundamental research, however, is the tive credit delivery system and appropriately- need for applied, site-specific research; farmers designed financial instruments will be required, need assistance in fine-tuning their cropping particularly to meet the needs of smaller/poorer practices to improve yields and avoid degra- households.'2 dation, not fundamental qualitative changes. Fertilizeravailability. To be useful, fertilizer must - Extension. If the results of research are to be of be available in a timely way. Poor experience use to farmers, a much improved extension with government control of the fertilizer sector system, which must not only deliver infor- led to a policy of gradual deregulation during mation to farmers but also convey information the 1980s which has improved supply sub- back to the research system on farmers' needs stantially. Problems persist, however, in the and constraints, is required. Efforts are supply of urea, for which a government para- underway to improve the effectiveness of the statal is the only source. The problems in urea Department of Agricultural Extension (DAE), but supply need to be addressed and must be much work remains to be done. Extension is an prevented from causing any back-tracking in area in which NGOs in particular might play a liberalization efforts. useful role; already, NGOs such as BRAC have made important contributions. * Fertilizer price. The need for higher and more appropriate fertilizer applications should not be Bangladesh's research and extension system, and interpreted as a call for the re-introduction of particularly research institutions such as BRRI and subsidies. Indeed, the current subsidies on urea BARI, have made important contributions to are probably having a harmrful effect by en- agricultural development in the past. What is couraging relatively more nitrogenous fertilizer required now is a re-orientation of research and to be used. Additions of nitrogen fertilizer extension efforts towards an increased focus on alone can give a short-term boost to yields, but site-specific conditions, on long-term research, only at the cost of further depleting other and on cropping systems and practices used by nutrients. farmers. It is important to ensure that both research and extension are responsive to farmers' The impact of efforts to diversify agricultural needs and constraints. Fertilizer recommen- production on the sustainability of intensives dations, for example, need to reflect site-specific agricultural practices is uncertain. Introduction of Environmental Economics Series 19 Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh crops with lower or different nutrient require- 1983-84 agricultural census showed that aman area ments can help slow or reverse nutrient depletion in particular had been under-estimated. Collection problems; in such cases, rice yields might and sampling procedures were revamped in light themselves improve. On the other hand, some of these results. The analysis in this report relies alterative crops, such as sunflower, are know to primarily on data collected after the census, and on ore-census data adlusted to correct for the observed be very demanding of soil nutrients. Their intro- pre-ces. duction into a rotation would probably not be as discrepancies. helpful from the perspective of sustainability. 7. BBS yield data are given in terms of metric tonnes Diversification can also help in other ways. (mt) of clean rice per acre. In this report, all yields Wheat, for example, has similar nutrient require- are quoted in terms of mt of paddy per hectare. The ments to rice so a rice-wheat rotation rnight recovery rate of dean rice from paddy is about 70 encounter the same nutrient deficiency problems percent. as a rice-rice system. By keeping soils dry for several months, however, some of the adverse & Paradoxically, local varieties appear to have expe- changes in soil physical and chemnical charac- rienced significant yield increases in the last fifteen teristics that result from long-term irrigation years (Table 1), probably due to spill-over effects might be avoided. Here too, more research is from improved management and input availability resulting from modern variety use. In particular, needed to determine both the nature and the inrasn aviablt of irrigation ha aloe extent of the sustainability impacts of different supplementaryvirrigationhoflaman. cropping systems. 9. National-level statistics (Figure 2, Panel C) provide Notes a long-term time series of input use, but do not distinguish between uses. Surveys which distin- 1. It is important to distinguish between observed guish between use on different crops, varieties, and yields, which measure output per unit land, and seasons were carried out in 1979-82 by the Inter- productivity, which measures the effort required to national Fertilizer Development Center (IFDC) and obtain that output. If input use is increasing, yields in 1989-90 by the BBS. However, differences in can increase even if productivity is declining. sample design make it risky to compare their results except at the national level, as in Figure 7. 2 Optimal fertilizer use is determined by the nece- ssary condition for profit maximization, that the 10. Similarly, although flood control and drainage marginal cost of fertilizer application equal its structures (FCD) can help growth by reducing or eli- marginal value product. minating the potential for damage to irrigation equipment and standing crops by floods, im- 3. In practice, however, the nature of available data properly designed FCD can lead to drainage means that the 'yield gap' that is generally reported problems which can adversely affect productivity. is that between current yields and maximum Flood control measures can also reduce or eliminate potential yields (YPX - Y0). the benefits of silt and nutrient deposition by flood- water. The extent of the benefits derived from such 4 Over-use of pesticide also led to health problems deposition is the subject of some debate, but there is and environmental pollution, as discussed below. very little hard evidence on the issue. 5. Even where yields are reported to be increasing, the 11. The simulations do not allow for the small observed reasons are instructive. Farmers in the Jessore- increases in local variety yields. They would make Kushtia area attributed yield increases to adoption only a very small contribution to growth since the of modem varieties, higher fertilizer inputs, and area planted to local varieties is declining rapidly. irrigation [Saunders, 1991]. These factors involve switching between systems rather than yield 12. This does not imply a need for subsidized credit. increases within a given system. Evidence from throughout the developing world indicates that the traditional supply-led approach 6. The accuracy of BBs data, particularly data collected to rural finance, which aims to provide 'cheap' prior to the 1983-84 agricultural census, has been credit to poor farmers, results in a system that by- questioned in many instances [Rashid, 1989; Boyce, passes the poor and that is financially unsustainable. 1985]. Comparison of BBS data with data from the 20 Environment Department Papers 4. Environmental Problems In addition to sustainability problems, concern has synthetic pyrethroids and moderately persistent been expressed that intensive agriculture harms organo-phosphates and carbamates and away the environment more generally. In particular, from highly persistent organo-chlorides. Use of the rapid increase in the use of pesticides is most organo-chlorines has been banned, and thought to pose two threats: production of DDT at Chittagong was halted in 1992? Heptachlor, which is registered for use on * adverse health effects for farm workers and sugarcane, is the sole exception to the ban on others exposed to pesticides; and organo-chlorines. * contamination of ground- and surface water, Even more so than with chemical fertilizers, it is harming downstream users of that water and difficult to obtain an accurate picture of the level damaging inland fisheries. of use on specific crops and how it has changed over time.3 Moreover, the threat of environmental This section examines the environmental damage damage depends not only on the quantity applied that might arise from intensive practices. The em- but also on the specific types of pesticides being phasis is again on the impact of the more intensive used and on their mode and timing of application. production practices in rice. Since environmental In this respect, the trend towards increasing use of damage results from the collective impact of granular pesticides is a positive one, since they are pesticides from numerous sources, however, use less likely to contaminate water than foliar pesti- of pesticides on other crops is also examined. cides and, because of their lower concentration of Differences in pesticide use across crops is of active ingredients, are often less toxic. Synthetic particular concern in the context of on-going pyrethroids pose particular problems: although efforts to diversify production away from rice.' their very low persistence makes them generally very safe for human use, they can be extremely Pesticide use in Bangladeshi agriculture toxic to fish even in small concentrations. Pesticide use grew rapidly in the early 1970s, Rice. About 70 percent of pesticides are used on following the introduction of modern rice vari- rice. Almost all pesticides used on rice are eties. Use fell dramatically in 1973-74, however, insecticides, although fungicides are occasionally following the halving of pesticide subsidies used against rice blast. Almost 90 percent of pesti- (which were removed entirely in 1978). Use has cide applications on rice in 1989-90, byvalue, were since increased again; sales of pesticides doubled on modem varieties. Usage is heaviest on boro, in the second half of the 1980s. Granular pesti- which received over 50 percent of pesticide cides have grown in importance, and currently applications on rice, by value, in 1989-90. Never- account for almost 80 percent of total pesticide theless, the amounts used per unit area and the use. Since granular pesticides contain lower con- total area affected are both relatively small. In centrations of active ingredients (ai), the quantity 1989-90, only about 10 to 20 percent of the area of active ingredients used has increased more planted to modern variety rice was treated with slowly than total quantity. There has also been a some kind of pesticide (compared to about 1 to 3 trend towards increasing use of non-persistent percent of the area planted to local varieties). Environmental Economics Series 21 Envirorunental and Natural Resource Degradation in Intensive Agriculture in Bangladesh Insecticide use on rice is mainly reactive rather [Kabir et al, 1994]. Other vegetables have lower than prophylactic. Although the initial Green rates of use, but still generally higher than on rice. Revolution package called for three scheduled It is common to spray cauliflower and cabbage 3 applications of pesticides, few farmers seem to to 4 times, for example. A large variety of pesti- have followed this practice since subsidies were cides are used, of varying hazard classes; phos- removed in the mid-1970s. All the farmers inter- phamidon, for example, is again widely used. viewed during field trips to the northwestern and Even low-toxicity pesticides can cause problems, southwestern regions of Bangladesh in the fall of however, at the high rates of use found on 1994, for example, used insecticides only upon vegetables. There are indications that resistance to detecting insect infestations in the fields. Farmers pesticides is well established among several generally could not identify the specific insecti- vegetable pests-in particular among pests that cides used; the type of insecticides, the dosage, attack eggplant and country beans. and the mode of application were generally recommended by local dealers based on the Other crops. Commercial crops such as tea, cotton, farmers' descriptions of the symptoms. Most jute, and sugar cane, have the longest history of farmers seemed aware of the safety precautions pesticide use. Although most generally cover required, and reported taking at least elementary relatively small areas, significant proportions are precautions such as wearing long-sleeved clothes, treated with pesticides. Of particular concern is covering the face with a scarf, and washing after the continued use of heptachlor, an organo- use. Again, the dealers appeared to be a primary chlorine, in sugar cane production. Sugar mills source of inforrnation on precautions. often provide heptachlor as in-kind credit to cane producers. Although use of heptachlor is only The three most commonly used insecticides are authorized on sugar cane, inevitably its use spills carbofuran, diazinon, and phosphamidon. over onto other crops. Several farmers inter- Phosphamidon is a foliar insecticide, while viewed in a sugar-growing area west of Bogra in carbofuran and diazinon are both granular. All November 1994 had used heptachlor on their three are relatively short-lived. Depending on the potato fields, for example. Saunders [1991] also formulation, carbofuran is classified as a class IA, reports use of heptachlor on wheat. IB, or II insecticide by the World Health Orga- nization (WHO), indicating that it is moderately to Main concerns extremely hazardous. Phosphamidon is classified in class IB (highly hazardous) and diazinon in Health problems. The toxicity of pesticides threat- class II (moderately hazardous). Carbofuran is ens the health of users. Health hazards can be also dangerous to fish. In 1989-92, carbofuran either acute, generally resulting from short-term accounted for about 13 percent of all pesticides exposure to relatively high doses, or chronic, re- sold, by active ingredient; this was a substantial sulting from prolonged low-level exposure. Use increase over its roughly 5 percent share in 1985. of toxic chemicals by illiterate farmers has led to Phospharnidon's share, on the other hand, concern. A study of rice farmers in the Philip- declined from about 37 percent in 1985 to about 20 pines, for example, found substantial health costs' percent in 1989-92. Diazinon, with about 29 associated with pesticide use. Indeed, the extent percent, has the largest share of the insecticide of health problems was sufficient to completely market-essentially unchanged since 1985. outweigh any benefits arising from pesticide use [Rola and Pingali, 19931. The level and nature of Vegetables. Use of insecticides on vegetables pesticide use in Bangladesh, however, differ follows a pattern almost diametrically opposed to significantly from those found in the Philippine that found in rice. Whereas rice is sprayed-if at study area, where heavy doses of pesticides were all-only 2 or 3 times a season, it is conmmon to applied on a prophylactic basis with few pre- spray vegetables such as eggplant and country cautions. The only farmers in Bangladesh who beans several times a week. A survey of eggplant might have levels of exposure to pesticide as high producers in the Jessore area showed a range of as-or perhaps even exceeding-those of Filipino application from 17 to 150 times for one crop cycle farmers are vegetable farmers. Awareness of the 22 Environment Department Papers Environmental Problems toxicity of pesticides appears to widespread farmers, and whether pesticide residues on vege- among Bangladeshi farmers. Indeed, pesticides tables pose a threat to the health of consumers. are colloqially referred to as bish, which means poison. As discussed above, farmers do take some Water pollution. Use of agro-chemicals can also precautions in using pesticides. result in health problems through pollution of drinking water by residues. Here the effects of No systematic studies of the health effects of pesti- fertilizer use are also a concern. Nitrates can be cides have been undertaken in Bangladesh. Anec- carcinogenic, for example, and can lead to blue dotal evidence from doctors indicates that cases of baby syndrome. The very shallow water table acute poisoning occasionally occur. The majority found in much of Bangladesh makes groundwater of these are associated with suicide attempts -the major source of drinking water for the rural (which is evidence of the awareness of pesticides' population-particularly vulnerable to conta- toxicity). There is no evidence of substantial mination. An important factor which needs to be chronic problems, but this is far from conclusive taken into consideration in any discussion of given the relatively short history of pesticide use water pollution in Bangladesh is the flushing in most areas, suggesting that cumulative expo- effect of floods and monsoon. There is some po- sure may not yet have reached dangerous levels, tential for chemical residues to accumulate during and thepossibility that pesticide-related problems the dry season, when use is heaviest and runoff are miis-diagnosed by health workers unfamiliar from agricultural land flows into beels and baors, with them. but residues generallywon't accumulate year after year. The threat of their reaching dangerous Heavy pesticide use on vegetables poses levels, therefore, is substantially lower than in particularly important dangers. Although rice mnany other countries. accounts for the greatest share of pesticide use, pesticide use is widely dispersed and doses are Until recently, little hard evidence existed of low. In vegetables, on the other hand, doses are actual levels of groundwater contamination. Tests high and applications frequent. Because vegetable of groundwater taken from village hand-pumps production is concentrated in a few areas and carried out by the National Minor Irrigation among a subset of farmers, the population expo- Development Project (NMIDP) have begun to fill sed is small but at high risk. Moreover, because of this gap. Traces of pesticides were found in only their management intensity, vegetables are often ten of 78 samples, despite the samples having grown close to the household, thus creating the been drawn from areas considered most at risk of potential for exposure of women and children and contamination. All the pesticides detected were of household animals. Vegetable plots are also longer-lived organo-chlorines; no traces of the often located next to ponds or waterways to moderately persistent organo-phosphates which facilitate irrigation, thus creating the potential for account for the bulk of pesticide use were found. contamination. Finally, vegetables tend to be Greater evidence was found of nitrate contami- sprayed heavily up to the time of harvest, and nation, but even the highest concentrations found then shipped directly to market with no waiting (20-30 gg/1) were below the WHO safe drinking period; moreover, many are consumed whole. water standard of 50 pg/l [de Zanger, 1995]. This creates a very significant potential for pesti- cide residues causing negative health effects on Degradation offisheries. Fish provide a substantial consumers. proportion of proteins consumed; the rural poor in particular depend on fish from open-access In light of the substantial adverse health effects capture fisheries.4 The output from capture fish- encountered among Filipino rice farmers, a careful eries appears to be declining, although the rate study of possible problems among Bangladeshi and extent of this decline are hard to determine rice farmers is warranted, even though the risk from available data, especially in light of the factors do appear to be lower. The greater prio- inherently high variability in fisheries output.5 rity, however, is to ascertain the extent and exact The decline in capture fisheries has been attri- causes of health problems among vegetable buted to a variety of factors, including over- Environmental Economnics Series 23 Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh fishing, modifications to water flows resulting use. BR23, which was first released in 1988, is from flood control programs, draining of beels for much less susceptible to brown plant hoppers and use in agriculture, and the effects of agricultural green leaf hoppers, for example, than the earlier and industrial pollution [Ali, 1990], but the BR11 variety which is currently widely used in relative importance of these factors cannot be aman. Second, and counteracting this first effect, determined from available data. Given the low intensification promotes pest populations, as overall level of use, however, pesticides appear discussed above. The net effect on the need for unlikely to have been a major cause of the decline. pesticides, therefore, is uncertain. Efforts to sti- There is some anecdotal evidence of fish kills re- mulate diversification add a further element of sulting from pesticide applications. These appear uncertainty, since alternative crops may entail to be related primarily to the use of highly toxic either higher or lower pesticide use and may pesticides of Indian origin, which are easily introduce new pesticides whose toxicity, available in border areas. Samples of fish col- persistence, and other characteristics differ from lected near Tangail by the Flood Action Plan those in current use. fisheries study found no organo-phosphate resi- dues but low levels of organo-chlorine pesticides Despite the apparent low level of current prob- [FAP 17, 1993]. The study was too limited, how- lems, there is room for improvement. A number ever, for general conclusions to be drawn from it. of relatively simple policies and reforms could reduce pesticide use further and make it safer. Rice-fish culture, in which fish are grown directly These same policies will also help to push future in the flooded rice fields, poses particular prob- growth away from the most damaging paths that lems. Rice-fish culture is said to have been pre- it could take. In general, the need for intervention valent in the past, and its small scale today is often is greater in the case of environmental problems blamed on the use of pesticides; however, this than in that of the natural resource problems conclusion is hard to reconcile with the small discussed in section 3. Most environmental proportion of total rice area which is treated with problems are externalities, in which the agents pesticides, particularly during the rainy season, causing the problems have no incentive to take when conditions are especially suitable for rice- those problems into account. Health problems do fish culture. Nevertheless, it does appear that have a direct impact on farmers using pesticides, pesticide use is incompatiblewith thevarious rice- but the link between pesticide use and health fish culture practices being introduced by NGOs problems-especiallychronichealthproblems-is such as CARE. often sufficiently unclear that farmners may not take it into account. Possibilitiesfor action In one important respect, Bangladesh has already With the exception of pesticide use on vegetables, adopted policies which help reduce the risk of the available evidence does not suggest the substantial environmental problems arising from existence of significant current problems resulting pesticide use in agriculture. By lifting the sub- from pesticide use in agriculture. There is sub- sidies that it paid towards the use of pesticides, stantial concern, however, that the extent of Bangladesh has avoided stimulating high levels of problems might increase as intensification con- pesticide use, as has happened in other Green tinues. Revolution countries. Countries such as Indonesia came to learn the hard way of the environmental Predicting future rates of pesticide use is difficult. and natural resource damage that such policies Since most pesticides are used on modem variety can cause [World Bank, 1994]. rice, total use will undoubtedly increase as con- version to modem varieties continues. Two oppo- Integrated pest management. Integrated Pest Mana- sing tendencies will be acting on this general gement (IPM) practices would allow the low trend, however. First, more recent modem rice current levels of pesticide use to be reduced sub- varieties have been bred for increased pest- stantially without adverse consequences for resistance, thus reducing the need for pesticide agriculture. One of the interesting aspects of IPM 24 Environment Departrnent Papers Environmental Problems is that it can be justified solely by on-site benefits, withthePesticideManufacturers' Association,but including (i) higher yields, resulting from less much work remains to be done in this field. damage being done to beneficial insects and a reduced risk of resistance building up; and (ii) Regulatory reform. Reform of the regulatory frame- lower costs. IPM has proven extremely successful work would help ensure that the pesticides that throughout South and Southeast Asia. The FAO are used are safe both for their users and for the has executed a pilot IPM project in Bangladesh as environment. The current system, which is based part of its Inter-Country Rice IPM Project. It on regulation of pesticides by brand name, has consists of a 30-hour farmer field school (imple- proven cumbersome and slow. In some cases, mented through the Department of Agricultural older, more toxic pesticides continue to be used Extension) in which IPM principles and techniques because newer, less toxic ones have not yet been appropriate to rice production are taught. It registered. This is the case of heptachlor, for appears to have been relatively successful in example. Alternatives that are far less toxic and reducing pesticide use among trained farmers, persistent are in widespread use around the with pesticide use falling by about 85 percent and world, but cannot be used in Bangladesh because yields increasing slightly [Ramaswamy, 19931. they have not been registered. Registration of Both UNDP and DANIDA are planning follow-on pesticides by brand name is also locking up scarce projects. CARE is also having substantial success research resources in essentially repetitive testing; with IPM techniques in its Integrated Rice-Fish reform would release these resources for some of (InterFish) Project. the more important research needs mentioned in the previous section. Ensuring that the costs and One problem with IPM is that being very time requirements for pesticide certification are knowledge-intensive, it is relatively expensive to reduced will become particularly important if extend to farmers. The FAO's pilot project requires current efforts to encourage diversification of 30 hours of training, divided into 8 sessions agricultural production bear fruit, since many of spread through the crop season. To date, about 77 the new crops will require different, specialized field schools have been carried out, with 50 chemnicals. Current registration efforts are also farmers each, for a total of about 3,500 farmers; focused entirely on assessing the effectiveness of CARE has trained an additional 6,000. Although pesticides, with very little attention being paid to the results have been very positive among trained safety aspects, either for users or the environment. farmers, the complexity of the knowledge re- quired for successful use of IPM practices has Regulatory reform efforts cannot be completely limited diffusion beyond the trained group. divorced from the situation in neighboring India, Given the cost of this extension, the relatively low where many highly toxic pesticides continue to be rate of pesticide use in rice, and the general produced and sold. Many Indian pesticides are weakness of the extension service, it may prove readily available in border areas of Bangladesh, beneficial to look for subsets of IPM which could where they are known to farmers for their be extended more simply. IRRI has had good "effectiveness", their high toxicity, and their rela- results, for example, with a rule of thumb which tively low price. Recent policy changes in India calls for refraining from spraying pesticides to will help reduce the price discrepancy between combat defoliating insects during the first 40 days India and Bangladesh and thus reduce the extent of the rice season [Carlson, 19941. More complete of the problem. IPM training could then be concentrated on areas in which pesticide use is highest. The other signi- Notes ficant limitation is that while IPM practices for rice are well-established, there are currently few avail- 1. In addition, significant environmental problems are able IPM practices for vegetables, in which the also often associated with the manufacture, trans- need to reduce pesticide is most urgent. Some portation, storage, and disposal of pesticides. For promnising research on IPM techniques for egg- lack of data, these are not discussed here. plants has been carried out by BARI in cooperation Environmental Economics Series 25 Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh 2 Stocks of DDT still exist, however, and its use is still approved in public health programs such as malaria eradication. Inevitably, some of these stocks have leaked out and been used in agriculture. The pro- blem should diminish in importance as existing stocks are exhausted. 3. The 1979-80 IFDC surveys, for example, report the percentage of farmers using pesticides and total use in quantity terms (lbs/acre), whereas the 1989-90 BBS surveys report the proportion of area on which pesticides are applied and total use in value terms (tk/acre). 4. Capture fisheries include fish which reproduce naturally and which are spread over the floodplain by the annual floods. 5. Since aquaculture output is increasing, however, the total production of fish may actually be increasing. The decline of capture fisheries has poverty impli- cations, however, since the poor have less access to pond fisheries than to capture fisheries. 26 Environment Department Papers 5. Conclusions Given the scarcity of land and the continued ping systems and practices used by farmers. The growth of population, there is no alternative but days of 'blanket' messages which are equally to continue intensifying agricultural production in applicable to all farmers are over; what farmers Bangladesh. As currently practiced, however, need is assistance in fine-tuning their cropping intensive agricultural production practices are practices to improve yields and avoid degra- degrading the soil resource base on which they dation, not fundamental qualitative changes. This depend, posing a substantial threat to their sus- effort will also require a much improved exten- tainability. Changes will be necessary to ensure sion system, which must not only deliver infor- that the achievement of self-sufficiency was not mation to farmers but also convey information just temporary. Environmentally, intensive agri- back to the research system on farmers' needs and cultural practices appear to pose much less of a constraints. Efforts are already underway to threat than is sometimes feared, with the impor- achieve these aims, but much remains to be done. tant exception of pesticide use on vegetables. For farmers to be able to increase the efficiency of The sustainability problems being experienced in their cropping practices, they will also need Bangladesh are not caused by distortions or reliable and timely access to inputs, including market failures inducing inappropriate behavior. irrigation, fertilizers, and credit. The liberalization Rather, they result primarily from the difficulties of irrigation in the late 1980s was largely of learning how to manage complex new agri- responsible for the sharp increase in growth in cultural production systems. The switch from that period. Liberalization of the fertilizer distri- local to modern varieties is much more than a bution system during the 1980s has resulted in change in type of seed. It is a change from a low- substantial improvements in fertilizer availability input, low-management system to a high-input, at the farm level. Problems persist, however, in high-management system. These sustainability the supply of urea; these problems need to be problems, therefore, are not amenable to tradi- addressed and must be prevented from resulting tional forms of government intervention through in back-tracking on liberalization efforts. manipulation of the price structure. Nor are they -appropriately addressed by rules and regulations Although the threat to the environment posed by on appropriate land use. Farmers have amply- intensive agricultural practices appears to be low, demonstrated that they have both the incentive several relatively simple interventions can help and the will to respond to degradation problems. ensure that they do not grow and may reduce What they need is a supporting infrastructure that them. Here too, what problems there are gene- will aid them in doing so. rally do not result from the impact of distorted policies. The most pressing need for action is in There is considerable scope for the research and vegetables, where- numerous factors indicate a extension service to assist farmers in developing high risk of damage to the health of farmers and appropriate responses to degradation problems. consumers and to the environment. Research is This will require a re-orientation of research urgently needed to determine the magnitude and efforts towards an increased focus on site-specific exact nature of problems, including threats to the conditions, on long-term research, and on crop- health of farmers and their families, of pollution of Environmental Economnics Series 27 Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh waterbodies, and of pesticide residues on food. Reform of pesticide regulation will help to ensure that pesticides in use are safe for users and the environment. The current registration system is slow, excessively concerned with pesticide effecti- veness rather than safety, and consumes valuable research resources. In many instances, it has resulted in older, more toxic pesticides continuing to be used because they are registered even though safer pesticides are available and in wide- spread use worldwide. 28 Environment Department Papers References ABEDIN, M.Z., and D. MUKHOPADHAY, 1990. BYERLEE, D., and A. SIDDIQ 1994. "Has the "Cropping System Based Fertilizer Green Revolution been Sustained? The Recommendations and Soil Fertility Quantitative Impact of the Seed-Fertilizer Investigations in Farmers' Fields." Fertilizer Revolution in Pakistan Revisited." World Demonstration and Distribution in Development, Vol.22 No.9, pp.1345-1361. Bangladesh, Field Document No.1. CARLSON, G.A., 1994. "A Review of the No Early TCP/BGD/8853. BARI On-Farm Research Spray Research Program at IRRI." Los Banios: Division and FAO. IRRI. ALAUDDIN, M., and C. TISDELL, 1991. The 'Green CASSMAN, K.G., S.K. DEDATrA, D.C. OLK, J. Revolution' and Economic Development: The ALCANTARA, M. SAMSON, J. DESCALSOTA, Process and its Impact in Bangladesh. London: and M. DIZON, forthcoming. "Yield Decline Macrnillan. and the Nitrogen Economy of Long-Term ALI, M. Youssouf, 1990. "Towards Sustainable Experiments on Continuous, Irrigated Rice Development: Fisheries Resource of Systems in the Tropics." Forthcoming in R. Bangladesh." Dhaka: International Union for Lal and B.A. Stewart, eds, Sustainable Conservation of Nature and Natural Management of Soils. Chelsea: CRC Press. Resources. DE DATTA, S.K., 1981. Principles and Practices of BHUIYAN, N.I., 1991. "Issues Concerning Rice Production. New York: John Wiley & Declining/Stagnating Productivity in Sons. Bangladesh Agriculture." Paper presented at FLOOD ACTION PLAN (FAP) 17,1993. the National Workshop on Risk Management "Investigation of Pesticide Residue Levels in in Bangladesh Agriculture. Joydebpur, Floodplain Fish in Bangladesh." Fisheries August 24-27. Studies and Pilot Project, Interim Report, Bouis, H.E., 1993. "Measuring the Sources of Annex B. Dhaka: FAP 17. Growth in Rice Yields: Are Growth Rates HANSON, R.G., 1994. "Soil Testing for Efficient Declining in Asia?" Food Research Institute Fertilizer Recommendations in Regional Studies, Vol.XXII No.3, pp.305-330. Research and Dervlopment." Prepared for BOYCE, J.K., 1985. "Agricultural Growth in Agency for Agriculture Research and Bangladesh, 1949-50 to 1980-81: A Review of Development and the World Bank. the Evidence." Economic and Political Weekly, INTERNATIONAL FERTILIZER DEVELOPMENT Vol.XX No.13, pp.318-331. CENTER (IFDC), 1982. "Agricultural BYERLEE, D., 1992. "Technical Change, Production, Fertilizer Use, and Equity Productivity, and Sustainability in Irrigated Considerations: Results and Analysis of Cropping Systems of South Asia: Emerging Farm Survey Data, 1979/80, Bangladesh." Issues in the Post-Green Revolution Era." Dhaka: IFDC. Journal of International Development, Vol.4 No.5, pp.477496. Environmental Economics Series 29 Environmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh IRRIGATION SUPPORT PROJECT FOR ASIA AND THE SIDHU, S.S., and C.A. BAANANTE, 1984. NEAR EAST (ISPAN), 1993. "Effects of Flood "Agricultural Production, Fertilizer Use, and Protection on the Fertility of Soils at the Equity Considerations: Results and Analysis Chandpur Irrigation Project." FAP 16 of Farm Survey Data, 1980-82, Bangladesh." Environmental Study. Dhaka: ISPAN. Dhaka: IFDC. ISLAM, S., 1990. 'The Decline of Soil Quality." In SOIL FERTILITY AND FERTILIZER MANAGEMENT S. Huq, A.A. Rabman, and G.R. Conway, eds, PROJECT (SFFP), 1994. "A Pilot Survey for the Environmental Aspects of Agricultural Soil Fertility and Fertilizer Management Development in Bangladesh. Dhaka: University Project: Kaliganj Thana." Volume 1: Main Press. Technical Report. Dhaka: SFFP. KABIR, K.H., F.M.A. RouF, M.E. BAKSH, N.A. TANDON, H.L.S., 1990. 'Where Rice Devours the MONDAL, M.A. HOSSAIN, M.A. SATTER, and Land." Ceres, No.126, pp.25-29. MS.M. ASADUZZAMAN, 1994. "Insecticide TASLIM, M.A., and F.U. AHMED, 1993. Land Usage Pattern on Vegetables at Farmers Level Leasing in Agriculture: A Study Based on a of Jessore Region: A Survey Report." Research Survey of Two Villages. Dhaka: Bureau of Report. Jessore Regional Agricultural Economic Research, University of Dhaka. Research Station. WORLD BANK, 1994. Indonesia: Environment and PINGALI, P. L., P.F. MOYA, and L.E. VELASCO, Development. Washington: World Bank. 1990. 'The Post-Green Revolution Blues in DE ZANGER, Frank, 1995. 'Water Quality Survey: Asian Rice Productions." IRRI Social Science Preliminary Results." Dhaka: National Division Papers, No.90-01. Minor Irrigation Development Project. RAMASWAMY, S., 1993. "A Review of IPM Activities and their Impacts During Boro Season 1993 in Bangladesh." FAO Intercountry Programme for Integrated Pest Management in Rice in South and Southeast Asia. GCP/RAS/145/NET. Dhaka: FAO. RASHID, H., 1989. "Land Use in Bangladesh: Selected Topics." In UNDP, Bangladesh Agriculture: Performance and Policies. Volume III: Land Water and Irrigation. Dhaka: UNDP. ROLA, A.C., and P.L. PINGALI, 1993. Pesticides, Productivity, and Farmers' Health: An Economic Assessment. Los Bafios: IRRI. ROSEGRANT, M.W., and R.E. EVENSON, 1993. "Agricultural Productivity Growth in Pakistan and India: A Comparative Analysis." Pakistan Development Review, Vol.32 No.4, pp.433-451. ROSEGRANT, M. W., and P.L.PINGALI, 1994. "Policy and Technology for Rice Productivity Growth in Asia." Journalfor International Development, Vol.6 No.6, pp.665-688. SAUNDERS, D.A., 1990. "Report of an On-Farm Survey: Dinajpur: Farmers' Practices and Problems, and their Implications." Monograph No.6. Dinajpur: BARI Wheat Research Centre. SAUNDERS, D.A., 1991. "Report of an On-Farm Survey: Jessore and Kushtia: Farmers' Practices and Perceptions." Monograph No.8. Dinajpur: BARI Wheat Research Centre. 30 Environment Department Papers Appendix Regression analysis of time trends of modem variety boro yields, 1979/80 to 1993/94 (t-statistics in parentheses) Division District Adi R2 DW Intercept Time Dummy 1 Dummy 2 Barisal Barisal 0.0363 2.51 3.51 0.0128 (37.2) (1.24) Patuakhali 0.418 2.66 4.16 -0.194 0.375 (12.1) (-3.45) (3.12) Chittagong Chittagong 0.344 2.39 4.00 -0.0619 (20.6) (-2.89) Chittagong HT 0.659 2.44 4.12 -0.101 0.114 (30.8) (-4.60) (2.44) Comilla -0.0704 2.86 3.96 0.00300 (41.0) (0.283) Noakhali 0.107 2.40 3.72 -0.0261 (25.6) (-1.64) Sylhet 0.457 2.14 3.47 -0.0521 (26.2) (-3.58) Dhaka Dhaka 0.213 1.83 4.45 -0.0332 (32.2) (-Z19) Faridpur 0.141 2.60 4.07 0.0463 (17.6) (1.82) Jamalpur 0.226 1.32 3.01 0.236 -0.259 (9.57) (2.45) (-0.36) Kishoreganj -0.0490 1.82 3.41 0.100 -0.107 (11.7) (1.12) (-1.05) Mymensingh 0.764 1.81 2.73 0.196 -0.172 (18.4) (4.32) (-3.33) Tangail 0.349 1.34 3.74 0.195 -0.251 (14.0) (2.39) (-269) Khulna Jessore 0.680 2.53 3.16 0.156 -o.28 (17.8) (5.36) (-3.67) Khulna 0.182 2.00 3.31 0.0560 -0.105 (21.9) (2.26) (-1.98) Kushtia 0.410 1.46 3.02 0.133 -0.256 (12.7) (3A1) (-3.09) Rajshahi Bogra 0.283 1.54 4.29 -0.0469 0.0518 (34.6) (-Z31) (1.20) Dinajpur 0.519 1.84 4.19 -0.0807 0.219 (27.4) (-3.22) (4.10) Pabna 0.110 1.99 3.60 0.208 -0.248 (9.66) (1.82) (-1.91) Rajshahi 0.203 205 4.11 -0.0458 0.164 (18.5) (-1.26) (2.12) Rangpur 0.360 2.24 3.99 -0.0453 0.105 (41.4) (-2.87) (3.12) Note: Dummy 1: slope change in 1983; Dummy 2: slope change in 1988 Source: computed from BBS data Environmental Economics Series 31 Envirornmental and Natural Resource Degradation in Intensive Agriculture in Bangladesh Regression analysis of time trends of modem variety aman yields, 1979180 to 1993/94 (t-statistics in parentheses) Division District Adi R2 DW Intercept Time Dummy 1 Dummy 2 Barisal Barisal -0.0878 276 278 0.0228 -0.0544 (16.6) (0.831) (-0.930) Patuakhali 0.488 2.732 3.37 -0.0546 (25.7) (-3.79) Chittagong Chittagong -0.0498 1.78 3.15 0.00929 (21.6) (0.579) Chittagong HT 0.274 2.44 3.43 -0.138 0.192 (12.2) (-1.61) (1.96) Comilla 0.420 211 Z51 0.0970 -0.123 (13.2) (3.13) (-1.86) Noakhali 0.0489 2.54 3.39 -0.2018 0.224 (8.52) (-1.65) (1.61) Sylhet 0.363 2.00 3.34 -0.188 0.249 (12.1) (-Z22) (258) Dhaka Dhaka 0.190 3.03 3.03 -0.0936 0.136 (11.5) (-1.17) (1.49) Jamalpur 0.244 2.25 3.07 -0.157 0.221 (8.62) (-1.44) (1.78) Kishoreganj 0.141 2.16 3.28 -0.0292 0.123 (16.03) (-0.870) (1.73) Mymensingh -0.0581 3.10 3.07 -0.0797 0.0971 (11.8) (-0.997) (1.07) Tangail 0.256 2.87 2.30 0.114 -0.102 (10.5) (1.70) (-133) Khulna Jessore 0.0560 1.63 3.42 -0.147 0.183 (10.2) (-1.43) (1.56) Khulna 0.691 1.60 2.91 0.0135 0.101 (24.2) (0.686) (2.41) Kushtia 0.187 1.48 3.51 -0.162 0.184 (15.1) (-Z28) (2.28) Rajshahi Bogra 0.132 1.92 3.46 -0.0982 0.133 (14.0) (-1.30) (1.55) Dinajpur -0.0136 1.44 3.30 -0.0387 0.0783 (18.5) (-1.32) (1.26) Pabna 0.488 1.74 2.58 0.0607 -0.0147 (16.6) (2.34) (-0.235) Rajshahi 0.537 1.86 2.80 -0.0479 0.121 (10.6) (-0.594) (1.32) Rangpur -0.00004 1.94 287 0.0843 -0.0870 (13.4) (1.28) (-1.16) Note: Dummy 1: slope change in 1983; Dummy 2: slope change in 1988 Source: computed from BBS data 32 Environment Departrnent Papers Environment Department The World Bank . 1818 H Street, N.W. Washington, D.C. 20433 202 473.3641 202 477 0565 FAX ® Printed on 100% post-consumer recycled paper