Poverty, public policies and the environment

ii I. e I rj1 i e 3.Iil §I .11 The paper has been prepared by N. Vijay Jagannathan, a Consultant to the Environment Department. The author is indebted to John English, Michael Mortimore, Jeremy Warford, R.A. Agunbiade, and David Pearce for their helpful comments. The author would also like to acknowledge the contribution of R.A. Agunbiade of the Nigerian Institute for Social and Economic Research, for having written one of the Nigerian case studies discussed in the paper. Departmental Working Papers are not formal publications of the World Bank. They present rough and tentative results of country and sector analysis or research, and are circulated to encourage discussion and comment. Citation and use of such a paper should, therefore, take account of its provisional character. The findings, interpretations and conclusions expressed in this paper are entirely those of the author and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. Because of the informal nature of this paper, and in order to present the results of research with the least possible delay, the typescript has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. * ii . In developing countries poverty and environmental deterioration are often visible in proximity to each other, and have led many to infer that a two-way causality exists between human and environmental degradation. The apparent association, however, draws away the attention of policy makers from more substantive causal factors. This paper suggests that the real factors causing environmental degradation may lie elsewhere, such as: in specific acts of public policy, through the expansion of consumption demand for natural resources by growing populations, and by the spread of urbanization. The poor are merely one of several actors who respond to changing incentive structures by altering their usage of renewable natural resources. The links between poverty and renewable natural resource degradation, the pal r argues, require evaluating the role of economic and institutional policies in altering labor and capital flows between and within regions. As economic and spatial integration of markets occur, several new marginal income earning opportunities become available in the informal sector of the economy, and the dependence of the poor on the natural resource base for livelihoods may actually get reduced. Three case studies from West Java (Indonesia), Ekiti-Akoko (S. Nigeria), and Gombe (N. Nigeria) are used to illustrate the linkages between poverty, public policies and renewable natural resource use. In all three, remote sensing imagery over two points of time provides basic data on land areas where vegetative cover appear to have undergone significant changes. A diagnosis of the causal factors has then been undertaken, and the paper demonstrates that the linkages between poverty and the environment have been largely influenced by economic and institutional policies. These policies have (a) shaped incentive structures and (b) improved social and physical infrastructure, through which natural resource usage has been affected. The poor, like the non- poor, have utilized opportunities brought about by the spatial integration of economic activities, sometimes to the detriment, and at other times to the benefit of long-term renewable natural resource usage. POVERTY. PUBLIC POLICIES AND TE ENVIRONMENT TABLZ OF CONTENTS IaM I. INTRODUCTION........... . II. THE MATURE OF POVERTY-ENVIRONMENT LINKAGES ............4 A. Poverty vs. Public Policies ...................... 4 B. A Conceptual Framework...........................5 C. Role of Public Policies ..........................7 D . Methodology............................ .8 III. CASE STUDY OF RABUPATEN SUABUNI......... .11 A. Changes in Land Use in the Baskoning Study......11 B. Rural Economic Changes. 1976-1986...............13 C. Role of Public Policies.........................14 D. Poverty and Environmental Degradation...........15 IV. TWO CASE STUDIES FROM NIGERIA........................18 A. Capabilities of the Physical System..............1 B. Changes in Land Use Between 1970s and 1980s..... .20 C. Implications for Poverty and Policy Analysis.....21 - Market Integration of Economic Activities......23 - Role of ADP Policies...........................24 - Role of Other Enabling Investments.............25 - Inequalities in Access to Resources............26 V. CONCLUDING RENMS...................................28 - Validity of the Study for Research and Operational Use.................................29 * 1 - POVERTY. PUBLIC POLICIES AND THE ENVIROWENT: SOME nELUTVMM-Y FINDINGS 1.1. Environmental degradation usually occurs when production and consumption activities of growing populations irreversibly weaken nature's recycling capabilities. These economic activities are also attributed to the development of markets, the advent of modern technology, and the spatial integration of inaccessible areas to market systems. In many developing countries, however, environmental degradation such as soil erosion, deforestation and pollution are most visible around poor settlements, leading some social scientists to highlight the direct links between poverty and the environment. 1.2. One has to recognize that both poverty and the environment are descriptions of states of human and natural resource attributes, and cannot be reduced to simple unidimensional cause-effect relationships. Apart from conceptual difficulties in modelling linkages, another handicap in many developing countries, is the absence of adequate or reliable data sets on poverty and environment characteristics. The challenge for operational research is exploring how circumstantial evidence and inductive logic can be used to explain the nature of interactions between the two states. This study is one such attempt, and evaluates the relative roles of poverty and public policies in changing the pattern of natural resource utilization with illustrations from Indonesia and Nigeria. 1.3. Perceptions of poverty-environment linkages in developing countries are usually traced back to the common denominator of high population growth tates, and rest on two premises: (a) Selective improvements in social indicators (notably life expectancy) have led to rapid increases in human populations of developing countries; (b) Economic systems in these countries have failed to create adequate income earning opportunities for the incremental working populations. 1.4. One could then argue, ceteris paribus, that poor individuals, in their search for food, would convert forests to farms, grow food on steep slopes and degrade marginal farmlands. Human history of several millenia indicate that people have cut trees, hunted animals and burned forests in their quest for food security. In poor economies the pressure of population would force agriculture on to more and more marginal lands, permanently impairing nature's regeneration capabilities, this argument would lead one to conclude. Finite natural resources would thus face increasing problems of negative externalities, and ecological systems would be irreversibly degraded in their physical capabilities. - 2 - 1.5. The linkages between poverty and environmental degradation are, however, not just governed by the physical limits of ecosystems, but rather, by the income strategies of the poor deslite the observable problems discussed in para 1.3. The reason is the penetration of monetization into remotest villages, with the spread of a "cash nexus" re-ordering economic and social values. Production and exchange through impersonal markets is strongly preferred to subsistence activities by poor producers and consumers alike. Although family decisions and mutual responsibilities may continue defining some aspects of household economic behavior, market exchange irreversibly supplants reciprocity and redistribution as the means of acquiring income and wealth (Dalton 1981). Under these circumstances, except for a small group of the ultra poor, most individuals and groups could be expected to strive for income security rather than just food security; cash being necessary to purchase non- traditional food and consumption items that reflect changing utility functions. A desire for income security has led to creative income earning efforts through a myriad of highly divisible, labor intensive, low skill vocations involving wage labor and self-employment in rural and urban informal markets (Jagannathan 1987, Pean 1989). 1.6. The availability of these (often unrecorded) income earning opportunities reduces the necessity of the poor to mine land and water resources close to their homes for survival. Equally important is whether the poor have the Rover to gain access or to use available natural resources. Poverty is very often visible as unequal access to land and capital because asset ownership in developing countries is inequitable. Moreover, monetization also appreciates land and water values to society, and what were common property rights begin to get privatized or nationalized. Under these circumstances, the poor are often unable to retain de facto access to productive land and water resources because of social and institutional factors, such as changing tenurial relations, laws and regulations, social and political relations both within and outside villages. Jodha (1986), for example, has documented how common property resources in drought-prone villages of India were gradually privatized following their increased economic values, leaving the poor de facto access to only marginally productive wastelands. 1.7. Apart from market incentive structures and institutional constraints, every developing country has attempted to improve the welfare of its citizens through social and economic policies. Before drawing any policy conclusions on poverty-environment linkages, one has to look beyond just the immediate physical environment, and correlate the largely individualistic survival strategies pursued by the poor to generate income and food security with: (i) Public policies designed to provide Basic Needs services to poor citizens, because of which economic expectations of the poor begin improving; (ii) Natural resource utilization by poor communities, taking into account spatial choices that are available for earning livelihoods between farming, off-farm activities and non-farm activities. - 3 - 1.8. The issue for research becomes disentangling complex causal processes, and explaining whether poverty, public policies, institutions or combinations of them cause changes in natural resource use. Analysis is further complicated because most developing countries do not have adequate time series data on either poverty-related or environmental variables. However, since 1972, remote sensing images have been regularly collected for most parts of the world, providing snapshot images of natural resource use at points of t1me. By matching aggregative changes in land use computed from these data sources with socio-economic data sets, a simple and quick method can be established to assess the nature of linkages between poverty, public policies and an area's land and water resource uses. 1.9. The object of this exercise is to provide one such practical method of understanding the linkages between poverty, public policies and the environment. The exercise has involved : (a) Identifying land areas where obvious change in natural resource attributes have occurred over a ten to fifteen year period in selected case studies, (b) Tracing the causal factors of observed land use changes to public policies and demographic factors by matching (a) with available socio-economic data sets. 1.10. At the first stage, gross changes in land use between two points of time (one from the 1970s and another from 1980s) are identified. These changes are then traced back to their likely causes in available data bases. If systemic causes could be attributed to either specific policies or economic variables identifiable from available data sets, the direct causal role of poverty would (by inference) be proportionately lower. Three subregions, Sukabumi (West Java), Gombe (Northern Nigeria), and Ekiti-Akoko (Southern Nigeria) were selected for analysis mainly because some data sets on the physical and economic systems were available. 1.11. The paper is divided into five sections. The second section lays out some conceptual ideas on the nature of poverty-environment linkages. Following this, the third and fourth sections describe the findings of the Wes,; Java and Nigerian case studies. The final section summarizes some general issues that may be of relevance in understanding the role of poverty, public policies and the environment in developing countries. -4- II. THE NATURZ OF POVERTY-EUMRUENT LMOAGS A. Poverty vs. Public Policies as Cause of Environmental Problems 2.1. Poverty and environment are descriptions of states: poverty describing the state of human beings in a physical and social environment under low, socially unacceptable, consumption and income levels; and environment describing the state of land, water and atmospheric resources under given or increasing consumption levels. 2.2. Poverty affects the use of land and water resources either directly or indirectly. The former occurs when the poor are forced by hunger and the lack of earning alternatives to mine land and water endowments of their community, and ensuing degradation either reduces or permanently impairs the productivity of those very resources. The latter takes place when public policies alter incentive structures and redirect capital and labor flows between sectors and regions. The poor, in these situations, are one of several actors who respond to economic opportunities by changing their established norms of using land and water resources. 2.3. The validity of the direct links between poverty and the environment can be questioned because: (i) The poor lack both the physical energy (because of nutritional deficiencies) and the capital equipment (because of lack of assets) to damage the environment in a significant way through production. At best, they could degrade open access resources in the periphery of their settlements during a search for "free" consumption goods. Given their low caloric and protein intakes, consumption of these non-food items would be fairly low. Evidence from India indicates that a slow deterioration of open access resources can still take place if a large number of the poor are involved in these gathering activities (Jodha op. cit.). (ii) The poor lack the knowledge beyond their own farming system experience, and also lack skills beyond supply of raw labor. If risk aversion is added to these attributes, their propensity to destroy natural resources is limited. One could then argue that it is only when the poor get transplanted to a new environment having a radically different farming system through public policy - as in the Brazilian Amazon (Brown and Stone 1989), or through open door migration of Sahelian and Savanna farmers to the forest zone of Ivory Coast - that unsustainable use of natural resources becomes observable. (iii) The poor, as rational individuals - even when they have the option of degrading the environment - would choose not to do so because of the likely damage to their limited future survival and income options. Kortimore's studies in Kano, Nigeria, for example, provides evidence that smallholder tree conservation is preferred to short- term income even under price and famine pressure. * 5 - 2.4. Enviromental degradation, according to this viewpoint, gets caused by: - The poor being misdirected through public policies, which are badly designed or poorly executed. Incomplete technological diffusion, poorly managed delivery systems and defective economic policies could be the actual causal factors; - The not-so-poor income maximizers utilizing labor of the poor for extractive production, - The ultra poor, for whom day to day survival is so desperate, that the future income is heavily discounted in favor of current income. However, individuals in this category are also likely to have very limited access to natural resources (see para 1.6 above). 2.5. In any case, irrespective of whether either poverty or defective public policies leads to resource degradation, rectifying the situation is complicated by poverty's indirect role in environmental degradation. Unlike the situation in industrialized countries, where natural resource degradation occurs as a by-product of identifiable wealth-generating production or consumption, in a developing country causes and effects get inextricably mixed up. A policy designed to avoid externality-induced costs is clearly preferable to a reactive policy toward the natural and human environment. B. A Conceptual_Framevork 2.6. Poverty has to be viewed from the perspective of both poor individuals and poor communities. There are two reasons why this is necessary. First, although strategies to earn incomes are largely individual-based microeconomic decisions, group norms and conventions play important roles in defining access to natural resources, as well as in determining individual attitudes toward common property resources. Second, a separate emphasis on poor communities also highlights the fact that in most developing countries, poverty is associated with the Nurksian low productivity and low capital formation aspects, and cannot be alleviated by redistributive policies alone. The linkages between poverty at both the individual level and at the community level with the immediate and wider environment are presented in Table 1. 2.7. The four groupings presented in Table 1 interact with each other in several ways. These are: * The interaction between a goor individual.and his or herimmediate environment. Poverty is manifested by low income and consumption levels, with suboptimal choices being made in consumption of basic needs (shelter, safe drinking water, education and health care). These suboptimal consumption decisions lower a person's physical quality of life, and become visible as degraded physical habitats and preventable environmental health risks. - 6 - Table 1 PovertV-Environmental Linkass State of being poor State of the environment Individual with Habitat covering living conditions low consumption and and access to amenities low expectations of future income Society or human Wider concept of habitat covering organizations of land, water and atmospheric resources poor communities of the area or subregion A second relation involves the individual with the land and water resources around his or her living area. Poverty is indicative of a lack of access to capital, technology and alternative sources of livelihood, and could also be associated with suboptimal individual production decisions, because of high private discount rates. Farmlands and pasturelands could get over-exploited, investments in soil conservation neglected, and crops could be grown on steep mountain slopes. One has to, however, view income- earning decisions of the poor keeping in perspective: - Additions to the economic resource bases of existing communities through capital inflow, through either public or private investments, - Rural-rural and rural-urban migration from capital-deficient settlements to areas experiencing economic growth through public and private investments that lead to improved income and employment opportunities. Communities of Poor Individuals Interact With Their Immediate Enironment. Social norms and conventions determine how communal open access resources (village wells, common lands, modern service facilities provided by the State) get utilized. These socially enforced rules determine attitudes toward conservation and waste disposal. For example, in most poor societies, recycling of agricultural and domestic waste products is widely institutionalized. Traditional housing materials in many countries, similarly, utilize locally available natural resources. Socially determined conventions and norms usually change slowly, and are often unable to adapt to changed circumstances caused by demographic and economic pressures. For example, in West Java, the extensive use of wood for housing is no longer a socially optimal *7- strategy because of widespread deforestation. Communities have resorted to illegal felling of trees from (poorly policed) protected forests, rather than rationalizing current socially wasteful methods of using a rapidly dwindling natural resource (See para 3.23). * Lastly, communities also determine access to productive resources such as land and water. Although most countries utilize a formalized legal system to define access to land and water resources, high transactions costs in utilizing the formal system result in the continued importance of traditional rights of usage. The interface between the formal and informal system often determines the incentives for sustained use of productive assets such as land. The Land Use Decree in Nigeria, for example, attempted to codify customary tenures on land in different parts of the country. Despite the well-intentioned provisions of the decree, some of the rent-seeking elite have used statutory tenure as a means of over-riding customary land rights of village communities (See para 4.27). C. Role of Public Policies 2.8. Public policies affect one or several of the above relationships. Some illustrations given below sum up the types of poverty reduction policies that directly affect poverty-environment linkages: - Some policies have been targeted at the individual poor producer or consumer; special credit programs, guaranteed producer support prices, food subsidies, targeted nutrition programs are some examples. These programs reduce the dependence of the poor on just the natural resource base of their immediate environment. A second group of more general policies have attempted to lift whole communities out of the poverty trap. Many of these social infrastructural interventions are like 'software', seeking to augment human capital of poor communities by spreading literacy, skill formation and preventive health care. - A third group of policies have addressed spatial deficiencies by improving the physical quality of life in communities, and by integrating rural areas with market centers. Investments in physical infrastructure are of this category, and have improved the 'hardware' through roads, permanent housing, community buildings (schools, health centers etc.), water supply, electricity, transport and communications. Investments of this category increase the mobility of labor and capital between locations. - A fourth set of policies have attempted to remove or reduce institutional constraints by specifically addressing the problem of access of the poor to productive resources. Land reforms, land re-distribution, input delivery systems, construction of neighborhood Farm Service Centers etc. are examples of such policies. - 8 - 2.9. All these poverty alleviation measures, together, are designed to improve access of poor individuals and communities to income earning opportunities by integrating their skills and work capabilities with the dynamic sectors of the economy. The policies, in theory, should reduce the direct dependence of the poor on local natural resource bases for subsistence. In practice, public investments and policies get affected by rent seeking behavior of the elite, and filter down to poor individuals and communities unevenly (Jagannathan op cit). Spatial and distributional equity have, consequently, rarely been maintained. 2.10. Major cities, economically advanced regions, and towns typically receive disproportionate shares of investments in terms of both software and hardware. The economic structure usually reinforces these trends by creating forward and backward multipliers in and around areas of population density because informal sector income and employment opportunities are usually concentrated around production and trading nodes. 2.11. The spatial dynamics of povercy ultimately determines the pattern of land and water degradation in even remote and physically isolated rural communities. Inter-relations between human population mobility, search for income security, public policies and environmental degradation are key to understanding poverty-environment linkages. 2.12. The consequences can be dramatic from country to country. In densely populated areas such as Java, very high densities (exceeding 700 persons per square kilometer even in rural areas) have been accompanied by the generation of a wide range of alternative income earning opportunities primarily in the tertiary sector (transportation, domestic service, trade and construction). With improved transport and communications, the rural poor are able to commute to towns for part-time or full-time work. In Nigeria, although similar vocations in the tertiary sector have been reported, the extent of spatial integration between rural and urban areas appears to be much less, resulting in less employment creation in rural areas. In both countries, public investments, notably in highway networks, inter-city transportation and urban development have greatly expanded population mobility of the poor. At the same time, the indirect effects of poverty on the environment are severe: negative externalities in the form of congestion, land and water pollution. 2.13. In the three preliminary studies, two large countries from Asia and Africa were chosen primarily because of data availability, and because they represent three distinct agro-climatic regions. West Java is typical of densely populated rice-growing areas of Asia. Gombe is a sample area from the Guinea Savanna belt of West Africa, while Ekiti-Akoko represents a sample area from the Tropical Rainforest Belt. D. Methodolony 2.14. The case studies identify changes in natural resource bases in selected areas by comparing temporal remote sensing imageries, and looks at the likely economic and demographic explanations by using available socio- economic data. The reasons for examining the consequences before understanding the causality were two-fold. First, there are as yet no - 9 - clear yardsticks of assessing qualitative changes in resource use in developing countries, although environmental literature is replete with general statements such as how forests are being encroached by land-hungry peasants, marginal farmlands depleted of their nutrients, river systems are being polluted by sewage from poor settlements, and so on. Corroborating these general statements with empirical data becomes difficult because information on land, forest and water resource characteristics is often unavailable. Unless a set of parameters of physical changes is worked out, it becomes difficult to translate general statements into specific operatienal concerns. Second, identifying the results before understanding the causes is more practical because socio-economic causality is a result of highly complex interactions among several independent explanatory variables. These variables affect resource use changes both directly (by altering incentive structures in the economic system and redirecting capital flows), and indirectly (by affecting movements of poor individuals in search of food and income security). It is more cost-effective to pinpoint and analyze the causes of ecological changes in identifiable locations, rather than disentangling complex causal processes first and then looking at the consequences on natural resource use. 2.15. Developing a method of identifying natural resource changes was, therefore, considered a pre-requisite for economic analysis. The ideal solution, if resources had permitted, would have been to design continuous monitoring systems for flow variables such as soil erosion rates, soil nutrient levels, areas deforested or exposed as degraded lands, b.o.d. and c.o.d. content of water, suspended particulate matter in the atmosphere etc. Such an effort requires large budgetary commitments, and is not affordable for most developing countries. A second-best solution was to develop cost-effective and simpler parameters of stock changes by computing temporal area statistics of land and water use. Temporal remote sensing imageries provided a relatively inexpensive, replicable and value-free basis of identifying aggregate stock changes in these natural resource endowments. 2.16. The methodology consisted of the following steps: * First, study areas ranging from 4,000 to 10,000 square kilometers were identified primarily on the basis of availability of supplementary data on natural resources of the subregion. * Second, LANDSAT images for two time periods (one from the mid 1970s and another from the mid 1980s) were used as basic reference material for visually identifying land and water use changes. Visual changes observed over a decade helped identify resource use changes, particularly in terms of land cleared of vegetative cover, growth of human settlements, areal farming changes etc. One could, consequently, restrict further research to spatially manageable micro-regions where noticeable physical changes had taken place. - 10 - * Third, causal factors in the identified micro-regions were analyzed by using socio-economic data. Specifically, available information on farm-nonfarm-urban informal sector linkages, data on rural-rural and rural-urban migration, public investments and institutional aspects (access to land, inequality in asset holdings) were combined together to set up preliminary hypotheses on the nature of poverty-environment linkages. 2.17. The selection of three study regions for this analysis from West Java (Indonesia) and Nigeria was largely determined by the availability of socio-economic and supplementary natural resources data. The studies have been reported separately in detail, and the following two sections are confined to discussing some of the findings of interest. - 11 - III. CASR STUDY OF KABUATEN SURABUMI (VEST JAVA) 3.1. A Dutch consulting company, Haskoning, had been commissioned by the West Javan Government to prepare an environmental profile for the province. Several data sets on natural resources and socio-economic variables were collected and analyzed in their report (Haskoning 1987). Of particular interest to this study are two remote sensing images for June 1976 and May 1986 for Kabupaten Sukabumi that were digitized and used to identify the land and water use changes in the Kabupaten or district. In this section an attempt is made to trace back their causes. A. Chantes-in Land Use in-the Haskoning Stud! 3.2. The Haskoning study has indicated a reduction in forest cover by about 27 per cent between 1976 and 1986 in Sukabumi, with the current forest cover being almost half the official forest statistics. According to the latter, there were 98,000 hectares of forests in 1986, while the remote sensing estimate was only 53,000 hectares. Overlaying sub-district (kecamatan) boundaries on deforested areas, ten kecamatans appear to have been severely affected by deforestation. The largest decreases have taken place in subdistricts that had (i) relatively low population densities, and (ii) improved road connections to the provincial highway system. Table 2 Changes in Forest Cover in Kabupaten Sukabumi (1976-1986) Chan&e f= forest Sg: Area (hectares) (a) Shrubland 19,864 38 (b) Mixed garden 7,133 14 (c) Estate 5,014 10 (d) Uncovered 1,999 4 (e) Built-up area 13,000 25 (f) Miscellaneous 5,230 9 Source: Haskoning (1987) 3.3. The interesting point revealed by Table 2 is that in about half the area of deforestation, land use has been changed to mixed gardens, estates and built-up areas, all obviously driven by market-related factors. These economic uses satisfy production or consumption requirements, and are sustainable with appropriate economic and institutional policies. The remaining area of apparent deforestation has occurred in the sparsely populated kecamatans, where conversion from forests to shrublands have taken place to satisfy consumptior demand of wood and other land based materials in West Java and Jakarta. * 12 - 3.4. Changes in land use in the district are different between the areas north and south of the river Cimandiri (which dissects the total area into two halves). North of Cimandiri river are some of the most densely populated subdistricts of Java, with population densities exceeding 900 persons per sq.km. In this area, the two remote sensing images suggest a systematic replacement of shrublands by rubber, tea, coconuts and clove estates/gardens on hill slopes with gradients between 2% and 15%. All these crops are important earners of foreign exchange, and reflect tl.e success of actively propagated policies designed to diversify the country's export base. With proper extension support, estates and mixed gardens could arrest soil erosion and simultaneously provide earning opportunities for the rural poor. Two other notable changes that appear in proximity of each other are the increases in water areas (because of public and private investments in irrigation), and increases in areas classified as bare lands (especially close to river beds). 3.5. South of the Cimandiri river, the spread of estates and mixed gardens has been more dispersed. ort of the southern kecamatans have both lower population densities and poorer land capabilities for farming. A significant change has been in the southwestern corner of the district, where a large tract of lowland forests (in the Cikepuh forest reserves) was reported to have been cleared extensively. 3.6 Process of environmental degradation: The Haskoning data suggests the conversion of forested areas to economic use following a sequence of changes. ESTATES -* BUILT-UP AREAS FOREST --* SHRUBLANDS -- -E STATES MIXED GARDENS -+ BARE LANDS The conversion of forest lands to shrubs appears as the first step. By 1976 itself, large tracts of what are ailt classified forests had been converted to other uses in the northern areas of the district. Between 1976 and 1986 all remaining sub-montane and coastal forests were severely degraded. The most serious damage to forests has taken place close to the highway system. Highway investments in the last decade have ensured that no area of the district is more than 20 kilometers from good roads, making all remaining forested areas susceptible to illegal felling of trees, mining or quarrying. 3.7. Next, in the sequence of changes appears to be the conversion of shrublands into economic uses, notably as estates or mixed gardens. This change is - in theory at any rate - both environmentally sustainable, and income augmenting. Dry land agricultural area, which is widespread in the unirrigated areas, does not appear to have changed substantially in area, although dry land paddy pr-duction increased through yield intensification (BPS 1986). 3.8. The only evidence of poverty-related unsustainable farming practices has occurred in abandoned/non-productive estates, which constitute about six per cent of the district's estate area. These properties have reportedly been invaded by landless cultivators and - 13 - displaced estate workers. Not having any legal titles to the lands, these persons have used ecologically destructive slash and burn methods of cultivation. Among the estates/garden crops, soil losses have been reported only in the tea estates during the immature phase of tree crop growth, until the crop canopy closed (Haskoning op cit). 3.9. The emergence of bare lands has generally occurred in areas where: - Farmers utilized very steep slopes for agriculture; * Land was mined for clay, gravel or limestone; - Abandoned estates were illegally farmed by shifting cultivators. 3.10. Finally, the Haskoning study indicates a very rapid increase in built-up areas, mainly through the spread of human settlements into adjacent shrublands, forests and farmlands. In aggregate terms, however, no decline in area under vegetative cover between 1976 and 1986 was visible in the densely populated subdistricts, perhaps because estates and mixed gardens extended vegetative cover on alreadv deforested lands. 3.11. More than deforestation, current environmental problems in this area relate to unregulated mining of mineral resources, clay, gravel and the over-exploitation of groundwater in the Sukabumi acquifer (Haskoning op cit). The study indicates that the economic system (and by inference not poverty) is the key determinant of land and water use. What is apparent is that market incentive structures, public investments in infrastructure, agricultural and macro-economic policies appear to be the main causal factors behind observed changes in land use. Each of these factors are discussed below, with the concluding paragraph evaluating the role of poverty in explaining environmental change in West Java. B. Rural Economic Changes, 1976-1986 3.12. Between 1976 and 1986, West Java witnessed increases in paddy production through yield intensification, spread of tree crop cultivation, and the widespread diffusion of commercialization. Investments in improving farm productivity, the buying and selling activities of State- supported agencies and private traders resulted in increased cash flows between farms and market centers. Favorable relative price movements of farm products and inputs, availability of critical inputs like seeds, fertilizers and pesticides, and the diffusion of appropriate technology allowed agricultural production of food crops and plantation crops to register significant gains. The growth of a commercialized rural economy, rural-urban integration of economic activities was made possible by social and physical infrastructural investments. A two-way flow of labor and capital between the urban and rural economies took place, with the following results: (i) Farmers who received higher cash incomes altered their consumption patterns, and consequently generated small multiplier effects in off-farm, labor-intensive supplementary vocations for landless' workers and marginal farmers; - 14 - (ii) Improved incentives for non-traditional off-farm activities linked to urban market demand, such as production of fish, meat, poultry and other products. Remenyi (1986), for example, has estimated a four-fold increase in milk supplies from West Java to urban areas between 1978 and 1984. (iii) Several low income, direct earning opportunities (becak driving, casual labor, hawking of merchandise, food preparation, etc.) were utilized in the urban informal sector of cities like Jakarta and Bandung. 3.13. Household budget studies indicate that these earning opportunities have altered the profile of consumption expenditure among the poor. On an average, rice (food) consumption expenditure declined in favor of non-food expenditure, while the population consuming more than 320 kg of rice equivalent increased between 1970 and 1976 (Hartmann 1985, quoting Sundrum and Booth). 3.14. Estimates indicate that more than one half of West Javan rural households are landless, and that about 60 per cent of landowning households owned less than 0.5 hectares of land (Hartmann 1985, Douglass 1987). Although no efforts had been made to reduce the inequitable distribution of land, West Java's poverty profile improved because the growth of income earning opportunities provided alternative sources livelihoods for these persons (Chandler 1984, Hartmann 1985). 3.15. The relative decline in poverty levels took place through occupational multiplicity among the poor; agricultural earnings were supplemented with cash incomes from other labor-intensive, low return activities in trade, construction, transport sectors. Rietveld's study shows that landless farmers working on farms derived 64 per cent of their incomes from non-agricultural activities. For farm households with holdings below 0.50 hectares, the percentage dropped to 55, and among households cxwning between 1 to 1.99 hectares of farmland, the percentage remained 40 per cent (Rietveld 1986). 3.16. The high proportion of non-agricultural income sources for even the poor suggests a widespread use of the market system for income generation and employment. Mining their immediate natural resource bases for survival or subsistence requirements appears to have been unnecessary. C. Role of Public Policies 3.17. Public policies of the Indonesian Government in the last decade have attempted to maximize rice and tree crops production through a combination of subsidies, easy credit, producer support prices, and infrastructural investments. Notable among them has been the BIMAS and INNAS rice programs. The expansion of agricultural production and incomes were possible because of three factors - availability of appropriate technology, favorable relative price movements of crops and inputs, availability of critical inputs such as fertilizers and water for farmers (Sjahrir 1986, Tjondoronigoro 1985). - 15 - 3.18. Other incentives such as the Village Development Assistance Pi,wgram provided subsidies for local development initiatives. Since the mid-1970s efforts have been made to reach out to special target groups, such as women traders, artisans, small home based industries through credit support, training, vocational guidance and extension (Hartmann op cit, Patten et al 1983, Tjokrowinoto 1982). 3.19. Imortance of enabling public investments: One of the most significant contributors to the integration of the rural and urban economies has been public investments in 'enabling' infrastructure, covering physical infrastructure, education and health, through which village populations have been able to seize income earning opportunities. The major facilitator of rural-urban economic linkages has been improvements in transport and communications brought about by the expansion of road networks and by the growth in public bus, truck, and motor cycle populations. By the mid-1980s, most villages were within a few kilometers access of public transportation. In a study of population mobility in the mid-1970s, Hugo (1981) found that the rapid expansion of cheap bus services within the province -reatly expanded circular migration and commuting between villages and sities. These 'temporary' migrants were left out of official statistics. According to his estimates, based on 14 villages in West Java, the temporary migration figures were roughly double permanent migration rates recorded in the 1971 Census. Another study based on SAE data appears to reinforce these arguments (Manning 1986). 3.20. Several other enabling investments have helped in the integration of rural communities with the market system. Health service facilities were greatly expanded, increasing life expectancy and reducing infant mortality rates. A drop in the country's total fertility rate of nearly one child per woman took place in the country between the late 1960s and the late 1970s (Sjahrir op cit). At the same time, however, the impact of past population growth has been so great that even with a successful birth control program, West Java's population is expected to increase by 40 per cent between 1985 and 2005 (Haskoning op cit). 3.21. Primary schorling facilities were also greatly expanded. By 1984 enrollment rates in primary schools reached 100 per cent, for the seven to twelve age group, making universal literacy an attainable goal (Sjahrir op cit). Improving literacy rates, higher life expectancy and smaller families have improved work force participation rates, and led to more rural-urban migration. D. Poverty and Environmental Degradation 3.22. The analysis of the economic system and public policies above suggests that the poor have by and large turned to the market system for their livelihoods. The Haskoning data also suggests that major changes in natural resource use can be traced to economic and policy variables. The informal sector, in both rural and urban areas has acted as a vast sponge, absorbing the growing population of job seekers. The burden of poverty is being most acutely felt in the densely populated settlements in terms of sub-standard housing, congestion, pollution, and health risks. Untreated solid and liquid wastes pollute local rivers, and downstream settlements (Haskoning op cit). - 16 - 3.23. Negative externalities of poor human settlements is felt by the surrounding natural resource base of an area. The Haskoning study has estimated fuel wood consumption in rural areas to range between 0.43 and 0.72 cubic meters per person, all of which gets harvested as free goods from village common lands. When the poor migrate to the city, in an attempt to minimize cash expenditures, they search for "free" goods from surrounding open access resources (including poorly policed public forests). If the relative pricing of substitutes to these free goods is unfavorable, the pressure on the latter would increase. A study by World Bank (1989) indicates that the urban retail price of kerosene in Indonesia has been increasing relative to fuelwood prices, and one could expect the pressures of illegal felling of trees to have continued unabated. 3.24. Apart from the pressures of poor human settlements, economic factors have led to excess demand for products from the area's natural resource base, covering the forest cover, sand, clay and gravel. For example, BPS data (1976) indicate'that housing structures in Indonesia by convention use wood and bamboo extensively, irrespective of income levels. If demand patterns remain unchanged, the high income elasticity of demand for wood products would continue to make tree felling lucrative. For West Java alone, the Haskoning study has estimated an annual demand of utility and construction wood of 1 million cubic meters; a figure substantially higher than local state forest production. Wood, as with most housing construction materials, has a low volume:value ratio, and supply areas with low transport costs expectedly face the highest pressure of felling. Poorly policed forests and river beds (for sand and gravel) in proximity to demand centers would, predictably, be the first to experience opportunistic mining of resources by private entrepreneurs, unless prohibitory regulations are strictly enforced. 3.25. Finally, another critical area of environmental concern is the very high population concentrations in erosion-sensitive watersheds of important rivers. The Haskoning study has indicated that eight district towns of the province are located in environmentally fragile areas, and most have rapidly growing populations. Tasikmalaya (an upland city) for example, has an annual population growth rate of over 3 per cent between 1971 and 1980. 3.26. Other examples of negative externalities reported in the Haskoning study are: - 6 - Increased pollution loads on rivers because of industrial and municipal discharges. - Intensification of mining of soil, gravel, clay and limestone in the vicinity because of urban housing and construction activities. These externalities, together with mining of acquifers in rural areas, misapplication of chemical fertilizers and pesticides appear to be the serious environmental problems of the area. The role of poverty in environmental degradation has thus been at best indirect. The direct causes of environmental degradation require policy corrections, innovations and physical investments. - 17 - 3.27. The solution lies in devising appropriate economic and institutional policies so that labor and capital flows are guided to environmentally optimal locations. Second, one can consider using subsidies on purely environmental grounds, so that the poor and the rich alike do not mine open access and poorly policed natural resources. For example, construction materials supplied from other Indonesian islands having adequate natural wealth could be subsidized for Javanese residents so that the mining of West Java's dwindling natural resources is reduced. Third, regulatory and policing institutions have to be greatly strengthened on common property resources. In sum, the answers to environmental degradation in Sukabumi have to be found in changing public policies, enforcing regulations and strengthening institutions. The role of poor indviduals and poor communities gets defined within the parameters set by thb e social and economic policies. - 18 - IV. TWO CASE STUDIES FROM NIGERIA 4.1. Two case studies from the Gombe and Ekiti-Akoko areas provide contrasting examples to the West Javan case study. These two areas fall within agro-ecologically distinct zones of Sub-Saharan Africa; Gombe is an example from the Savanna belt, while Ekiti-Akoko lies in the tropical rainforest zone. The choice of the two areas was again, largely dictated by the relatively better availability of physical and socioeconomic data because of their coverage by World Bank-assisted Agricultural Development Projects (ADPs). 4.2. Unlike the West Java study, where the Haskoning data base provided an invaluable foundation for analysis, the Nigerian studies required searching available data sets for relevant information on physical and socio-economic characteristics. In both Gombe and Ekiti Akoko, remote sensing images from the 1970s and 1980s had to be visually examined and related to other available maps and data bases, so that gross changes in land and water use in the intervening period could be evaluated. 4.3. Changes in the physical system were measured by overlaying maps showing land use and soil capabilities on remote sensing images of two time periods. The manual overlay method was not as technically elegant as the digital overlays used by the Haskoning study for West Java. One has to however, balance technical sophistication with the reliability of available data sets. Compared to Java, Nigerian data on both the physical and socio- economic systems are considerably weaker in accuracy and coverage. For example, no reliable population estimates are available in the country since the 1950s; there are no agricultural census data bases; cadastral surveys of land are virtually non-existent, and information on soils, slopes and land rights are subject to gross generalizations. There are serious risks of compounding existing errors in data bases by attempting ambitious digital exercises without adequate ground truthing. In this desk study the object has been only to test the validity of the method: of deriving generalizations on gross changes in the physical system as a surrogate for (unavailable) data detailing land and water use changes. A. Capabilities of the Physical System 4.4. An assessment of land capabilities was made from existing data sets in the two study areas. In Gombe, information on the physical system was obtained from the Bauchi State Atlas. The Atlas has computed land capabilities by relating data sets on soils (productivity, erosivity, slopes and suitability), water availability (rainfall, acquifers, water bodies), climate and vegetation with each other to qualitatively classify agro-ecological potentials in four major categories. For Ekiti-Akoko, information from maps used during the preparation of the Agricultural Development Project was divided into three categories: - The well-drained Ondo soils, with slopes less than 7 per cent have been classified as the 'best suited lands'. - 19 - - The medium capability lands are the Iwo soils, which have coarser textures and more frequent rock outcrops. * Lands with low capabilities covered all areas with slopes greater than 7 per cent, where bare soils are prone to severe erosion risks because of the intensity of run-off. Risks of soil erosion were common to both Gombe and Ekiti-Akoko, but water availability was a binding constraint only in the former study area. To ensure comparability between the two study areas, the four-fold classification of land potentials in Gombe was reduced to three groupings; moderately good farming lands, lands with generally inferior potentials, and lands with high erosion risks. Table 3 contrasts the land capabilities of the two study areas using this classification. Table 3 Land Classification in Gobe and Ekiti Akoko Gomba Ekiti Akoko Land with good potential (%) 30 48 Land with generally inferior potential (t) 33 16 Erodible lands (%) 37 36 Approximate area sampled (sq kms) 6400 3600 §ourcg: Bauchi State Atlas and IBRD map 14232 4.5. The land capability classification above has to be interpreted only as a XaLL distribution of good, average and erodible farmlands in the two study areas. Comparing across the two study regions is clearly inappropriate because the groupings are uniquely specific to the respective ecological zones. Lands with 'good potentials' in r.kiti-Akoko, for example, can support larger farming populations than similarly classified lands in Gombe. Perhaps the only common feature is the roughly equal proportion of lands subjected to erosion risks in Gombe and Ekiti-Akoko, although erosion-prone lands in Gombe are often interspersed among 'best suited' alluvial farmlands of the Gombe-Kumo belt. 4.6. By overlaying available data on population and human settlements on land use potentials, one observes, just as in Sukabumi, that agricultural and land clearing activities have occurred on better suited lands. In the Ekiti-Akoko area, higher population densities and clearing of vegetation were observable on the better endowed Ondo soils having less than 7% slopes. In Gombe, land clearing has taken place on the relatively good soils between Gombe and Kumo, and on the heavy clay soils along riverine tracts. - 20 - B. Chanmes in Land Use Between the 1970s and 1980s 4.7. Visual interpretations of vegetative cover changes were facilitated by contrasting false colors in the remote sensing imageries. In Ekiti-Akoko, sharply contrasting colors of vegetative and tree cover (in red), cleared land (in light green) and human settlements (turquoise blue) were visible. In Gombe, the spread of fadama cultivation (red), reduction in uncultivated tracts of clay soils (dark browni and tertiary sandstones of the low farming potential areas of the Kerri Kerri plateau (light brown) could also be visually identified. These visual interpretations were verified with ground truth from several sources, including a Nigerian economic geographer, ADP data sets, Bauchi State Atlas, and the ODA study of the land resources of Northern Nigeria conducted in 1971. 4.8. The following were the significant findings of the study: (i) Major changes in land use occurred on the periphery of larger towns, along roads, and at major highway intersections. In Ekiti- Akoko a ribbon pattern of land clearing is very obvious: all highways in the project area can be visually identified in the 1987 LANDSAT imagery. In Gombe, similar road networks could be identified in a ten kilometer radius around Gombe town. Public investments in roads appear to have been major contributors to the observed changes in land utilization patterns in both study areas. (ii) The clearing of lands for farming has generally taken place on better suited lands (as defined in Table 3). In the Ekiti-Akoko image for 1987, lands not connected to the highway system experienced very little changes in red color (vegetative cover), and white signatures (usually a sign of bare or degraded lands) were not observable. In the Gombe image, the most significant change was the diffusion of farming in uncultivated areas with good farming potentials. Visual examination in Gombe could not however, distinguish between bare lands, human settlements with thatched roofs and fallow farmlands. Overall, both images suggest that food crop farming has spread to better suited and accessible agricultural lands. (iii) Land clearing in Gme took place in areas that were earlier inaccessible (in the southwest), and in the heavy clay soils zone. Available (although scanty) socioeconomic data suggest that extensive farming spread into the areas serviced by new roads. ADP studies indicate that there has been a widespread diffusion of labor-saving oxen cultivation over these areas, which account for the significant areal increase observable. (iv) In Ekiti-Akoko, most cleared lands appear around existing settlements and road networks, with some of the isolated areas actually experiencing re-vegetation. Changes in land utilization in areas adjacent to large towns and highway networks are indicative of the spreading commercial farming of food crops (through crop switching from tree crops). Again, as in Gombe, ADP data do not indicate any significant improvements in average - 21 - productivities of food crops grown in these areas. Another significant change has been the virtual disappearance of the Eda and Agbe reserve forests, both located close to areas of high population density, in the 1987 image. (v) In Gombe the expansion of irrigated fadama agriculture on river valley bottomlands-was clearly visible. Large tracts of land below the Dadinkowa dam on the river Gongola had been used for irrigated vegetable cultivation. In the Ekiti-Akoko area, perhaps because of its location in the tropical rainforest belt, no analogous changes were visible in land use along rivers and water bodies. (vi) The growth in human settlements was confined to existing towns and villages in both study areas. Increased flows of public investments into social infrastructure (education, health, water supply) perhaps increased the attractiveness of towns for rural residents. (vii) In both areas vegetative cover was cleared around urban centers. In the Ekiti-Akoko area, major changes have taken place around all five Local Government Area headquarters, and in other towns located on the highway network. In the Gombe area, population density and settlement sizes were considerably smaller, and visible changes to a comparable degree were visible only around the largest town of Gombe. Table 4 summarizes some of the principal findings. C. Implications of the Visual Stud! for Poverty and Policy Analyais 4.9. The visual examination of remote sensing images has identified micro-regions that have exhibited systemic changes in land use. Specifically, the following areas appear to fit a consistent spatial pattern of change: Gonbe area - A periphery of ten kilometer radius around Gombe town, where maximum changes in land use appear to have taken place between 1976 and 1987 - Heavy clay soils to the west of Kumo, which have been brought under cultivation with new road investments - River valley bottomlands and fadama lands south of the Dadinkowa dam, where irrigated farming has diffused - Kerri Kerri plateau fringes to the west of Pindiga, where land degradation appears to have taken place - 22 - Table -4 Chanjes in Land Use Around Principal Towns zM Area of identifiable 'khan..qjince122 change (sq kilometers) (a) Mkiti-Akoko Ikare-Oka Oka- Erushu-Ogbabi belt 66 85 Arigidi- Osin belt 31 96 Ado Ekiti 25 75 Omuo 7 73 Ikole 13 81 Other towns on federal roads 17 74 Remaining towns 22 94 Other areas 30 95 (b) Gombe area % change since 1976 Gombe town 13 233 Source: LANDSAT imagery Ekiti-Akoko area - Urban peripheries of Ado Ekiti, Ikole, Ikare, Omuo, which are areas of greatest land use changes - Ribbon development around highways, particularly between Ifaki- Ikole-Omuo. - Periphery of the reservoir on river Osse - Disappearance of forests at the Eda and Agbe forest reserves 4.10. Changes in land and water use appear to be closely related to human settlements, economic policies and public investments. In West Java, many of the available socio-economic studies and reports could be directly related to the observed physical changes in land and water uses. By contrast, in both Gombe and Ekiti-Akoko village-level studies and socio- economic data bases were practically non-existent. The lack of any - 23 - reliable Census for more than two decades further weakened the analysis. However, limited available data suggest that changes in natural resource uses, just as in West Java, appear to have been influenced by market forces (increase in food prices, changes in relative prices between food crops and tree crops), by infrastructural investments (road networks, health and social infrastructure), and specific ADP policies, such as fertilizer subsidies, location of farm service centers, investments in rural water supplies etc.). 4.11. The weight of visible evidence points to the importance of incentive structures. institutions and public policiep in explaining changes in natural resource usage. Four sets of factors could be used, just as in West Java, to explain the observed changes in land and water use. These are: (i) The degree of market integration of economic activities; (ii) ADP policies designed to improve food production; (iii) Investments in basic needs and infrastructure; (iv) Prevailing patterns of inequality of income and access to land. (i) Market Integration of Economic Activities 4.12. The rural Nigerian economy has a long history of responding to favorable market signals, with areas under cocoa, cotton, groundnuts expanding rapidly when farmgate prices were remunerative (Bates 1981). In the post-SAL period, improvements in terms of trade for food crops has led to areal expansions of maize, cassava, rice and yams. Despite this trend, unlike West Java, rural income multipliers for the poor have been fairly low. Haggblade and Hazell (1987) discuss the causes, and a summary of their arguments has validity for both study areas. 4.13. First, backward processing linkages in the rural nonf4rm sector are very low. Fertilizers, agricultural implements and construction materials -the main generators of linkage effects in West Java - all had lower diffusion rates because there had been no equivalent of the Asian "green revolution". In West Java, by contrast, wetland paddy, tree crop production and input consumption provided major backward and forward processing linkages in the rural sector. Moreover, past policies of distributing inputs and farming implements through centralized parastatals and Farm Service Centers inhibited microenterprise development in rural areas. 4.14. Second, forward processing linkages in rural areas were also limited despite the substantial processing requirements of traditional food crops such as cassava and maize that have benefitted from price increases in the post-SAL period. ADP food processing studies (AMDEC 1984) indicate that manual techniques employed by village women to process cassava or maize produce outputs were too time-intensive to generate adequate marketable surpluses. The lack of satisfactory storage infrastructure and high transport costs in rural areas have further prevented rural household food processing units capturing urban market shares (Durojaiye and Aihonsu 1988). - 24 - 4.15. An ADP study of gari processing in Omuo town (Ekiti-Akoko) serves as a useful illustration of why towns, rather than villages have benefitted from forward processing linkages. Producers informally shared the costs of investing in a time-saving, but expensive poitable motorized grater for processing cassava. By collectively purchasing raw cassava and other inputs from local wholesale traders, risks of irregular supply have been minimized, and many women microentrepreneurs have been able to successfully compete against bulk producers of gari (AMDEC 1984). 4.16. Third, consumption linkages in demand for rural services such as construction, transport and durables depend on rural economic growth rates. In an economy where most forward and backward linkages have been restricted to urban areas, one could expect consumption linkages to be confined to urban and peri-urban areas as well. 4.17. Inadequate generation of non-farm incomes in rural areas has been mirrored by substantial employment multipliers being observed in Nigerian urban and peri-urban areas. Ukwu (1982) has estimated tertiary sector (distribution) activities to account for 24 per cent of Gross Domestic Product in Nigeria in 1982, and to employ about 15 per cent of the work force. He estimates double as many traders in urban areas as in rural areas. Together with the labor requirements of the construction and other services sector, urban and semi-urban areas are locations with better income earning opportunities. Unlike West Java, where the rural and urban economies became spatially integrated through private and public investments, in areas like Gombe and Ekiti-Akoko, the success of rural development efforts could be expected to depend largely on the nature and direction of public investments, a topic that will be discussed next. (11) Role of ADP Policies 4.18. Between 1972 and 1986 both Gombe and Ekiti-Akoko received large public investments through the Agricultural Development Projects (ADPs), designed to promote food production and rural development. These resource inflows complemented other nation-wide physical investments in transport, communications and in social infrastructure (education, health, water supply). 4.19. In both Gombe and Ekiti-Akoko, the ADPs were designed to correct some of the rural-urban imbalances discussed above, and disseminate the benefits of modern technology to households with small agricultural holdings. Increases in food crop production (observable as changes in land use in the LANDSAT images and from some farm-level studies), notably in maize, attest to successes in the programs. The evidence in both study areas suggest, however, that area expansions were not accompanied by yield intensification. There is also no evidence of the growth of occupational multiplicity, non-farm incomes and circular migration as in West Java. 4.20. The lack of productivity gains in Gombe is evident despite increased consumption of chemical fertilizers in maize cultivation during the last fifteen years. In Ekiti-Akoko, by contrast, the total quantity of fertilizers sold was barely a fifth of the appraised target because they were not considered suitable for widely grown staple root crops (Agunbiade 1988). - 25 - 4.21. The experience of diffusing other farm input has been similarly mixed. In Gombe, oxen cultivation became widespread, although leading to extensive rather than intensive agriculture by affluent farmers (Clough and Williams 1987). In Ekiti-Akoko, the Advanced Services Package and Basic Services Package had been designed to diffuse minimum tillage practices and introduce scientific methods into crop production. The use of herbicides, insecticides, better quality seeds and farm machinery were however, far below appraised estimates (APMEPU 1988). 4.22. Another set of problems can be attributed to organizational and management factors. A study of farm mechanization in Ondo state by Faborode (1984) serves as a useful illustration of why state-sponsored rural development programs did not galvanize agricultural development. In the ntire state of Ondo (including the Ekiti-Akoko area) there were only 218 tractors in 1983, of which 96 per cent were owned by government farms. These tractors had to service about 190,000 hectares of land under food crops in a fairly rugged terrain. Poor operations and maintenance of these machines further diluted their effectiveness - about 23 per cent were not working because of either mechanical defects or lack of spare parts. Farmers interviewed by the author complained of poor quality of service, frequent breakdown of machinery and problems of timeliness in machine availability. Tractor hiring services were confined to areas in the vicinity of the Farm Service Centers having good road access. (iii) Role of Other Enabling Investments 4.23. Infrastructural projects under the ADPs and other nationally- sponsored projects had several other impacts on the project areas. In both Gombe and Ekiti-Akoko, road access was improved, but investments had to be spread relatively thinly across when compared to relatively compact, densely populated areas like Sukabumi. In Gonbe, scarcity of drinking water has been a major problem in many settlements, and efforts have been made to provide a minimum water supply infrastructure (WARDROP 1986). The borehole program could not be extended to the heavy clay soils opened up by animal traction to agriculture, and few human settlements are visible close to these areas of good farming potentials. In Ekiti-Akoko drinking water supply has not been a problem, although the rural areas have been inadequately serviced by safe drinking water sources (Ondo Survey 1986). 4.24. Diffusion of primary schools: The most notable achievement of the last two decades has been the spread of primary schools in rural areas under the Universal Primary Education program, and a parallel growth in demand for schooling even among poor farm families. In Ondo state free education was introduced by 1979 itself, and school enrollments rose from about 580,000 in 1979-80 to over 700,000 in 1984-85. In Bauchi state, progress was not as impressive, but has been around 350,000 in 1984-85 (FOS 1985). 4.25. This expansion of schooling has led to tremendous pressures on institutions for higher learning in the cities. Only between a fifth and a tenth of applicants could be admitted to Universities. For the others, literacy has led to increased expectations, without a corresponding growth of farm and non-farm income opportunities in rural areas. Growing populations of school leavers are drawn to cities in search of income and employment in the urban informal sector. * 26 - (iv) Inequalities in Access to Resources 4.26. Unlike West Java, the rural areas of both Gombe and Ekiti-Akoko had fewer opportunities for the poor to earn incomes outside the farm sector, unless they migrated to urban areas. Within the farm sector, their utilization of natural resources would depend on institutional factors that determined access to land. The few available studies suggest that inequalities in land holdings have been noticeable in both the study areas. In Gombe, over a third of farms were reported to be less than two hectares in size. Another study of income distribution indicated the land:man ratio for the poorest households to be 0.31 hectares. The Staff Appraisal ReporL for Ekiti-Akoko estimated average farm size to be about 2.5 hectares, while Faborode's estimate of average arable farm size in Ondo state was between 0.24 to 0.4 hectares (Faborode op cit). In both areas, just as in West Java, a large proportion of households, had access to to limited productive farmlands. 4.27. Unequal access to land, and small size of holdings for a majority of households have led some authors to conclude that the benefits of public investments in agricultural infrastructure have not spread equitably across the rural population. In the Gombe area, the growth of extensive farming is reported to have been mainly financed and organized by the affluent elite (Beckman 1987). In other parts of Nigeria too, the vigorous expansion of food crops near urban areas have often been attributed to the entrepreneurial initiatives of public servants, traders and large landowners. Berry (1987) and Ross (1987) have expressed their concern that the Land Use Decree of 1978 is simply being used by the elite to appropriate valuable peri-urban lands for commercial agriculture with the help of State power. The Kaduna Land investigation for example, revealed that land accumulation by the affluent had taken place by manipulating the Decree (Watts 1987). 4.28. The issue of policy concern is whether these actions by affluent members of the community are sustainable in the long run, or whether they merely represent a short-term rent seeking strategy. If the new breed of farmers who lease in land with the backing of State power are not confident of retaining their access rights, fragile topsoils could be mined for short-term profits. The long term consequences could be the neglect long- gestation soil conservation measures essential to maintain fertility in farms where annual crops are being grown. In the peripheral areas surrounding urban settlements of Ekiti-Akoko, where extensive clearing of tree cover has occurred, such a trend could lead to serious problems of soil erosion. The question whether the social groups with traditional rights of usage have willingly participated in thc leasing in and leasing out of lands becomes critical in evaluating the sustainability of food crop production in these identified microregions. 4.29. Without further empirical support one can only speculate on the observed pattern of land use changes in rural areas. Further research is required to identify which of two opposite trends are dominating: - 27 - * First, because of market integration and attractive prices being obtained by food crops, the expansion of food crop farming in both the urban peripheries (Ekiti-Akoko), and in the heavy clay soils (Gombe) could be improving employment and income-earning prospects for the poor. Profits could be plowed back for more intensive and sustainable farming (as practiced in Kano) with minimum land degradation. * Second, because a (short term) rent-maximizing elite have utilized access privileges to farm these areas, serious environmental risks (soil erosion in Ekiti-Akoko) or sustained public commitments of resources (in maintaining road networks and fertilizers subsidies in Gombe) impose escalating liabilities on the exchequer, without improving farm productivity or increasing off-farm income opportunities. 4.30. The role of the poor in natural resource use would depend on which of these two trends dominate. If the first dominates, poor individuals have one more alternative, as in West Java, (apart from the choice of rural-urban migration) to subsistence farming in their villages. However, if the second dominates, growing populations of the poor will have virtually no alternative to subsistence farming. The poor and the rent seeking elite would together irreversibly damage their fragile natural resource base. LANDSAT images have helped identify the regions in both Gombe and Ekiti-Akoko requiring close monitoring to determine which of these two trends dominate. In areas where resource degradation is taking place, policies and plans to comprehensively manage the natural resource base are necessary. - 28 - V. CONCLUDING REMARKS 5.1. Natural resources data on soils, topography, vegetation, water etc. have been often utilized to assess the economic potentials of land areas. This study has demonstrated that from a policy perspective, it is equally important to understand how macro-economic variables and poverty determine land and water resources usages by communities. Land use changes over two points of time were identified, and then causality traced back to relevant economic factors. Specific investments by both the public and private sector, institutional factors, prevailing living and working conditions - all, to varying degrees - contributed to observed changes in natural resource use in the study areas. 5.2. In both West Java and Nigeria, major observable changes in the physical system since the 1970s such as, conversion of tree crop farms to food crop farms, clearing the bush for farming, disappearance of forest reserves and the expansion of human settlements, could be traced back to stimuli from economic incentives, rather than to just poverty-induced exactions. The economic stimuli have been influenced by several factors: the structure of product and factor markets, policies that encouraged private investments in agriculture, and the size and direction of public investments. 5.3. In rural areas the direct link between poverty and environmental degradation was difficult to establish, although in densely populated areas of West Java, the growth of substandard squatter settlements, land and water pollution, were clearly identifiable indirect links between poverty and the physical environment. In Nigeria, the direct links between poverty and resource degradation requires further ground-level research. 5.4. A study of three distinct agro-ecological regions from Indonesia and Nigeria suggests that: * It was possible in both countries, to relate changes in spatial areas of land and water use to specific economic decisions taken by individuals and the State. Public policies and institutions appear to have played critical roles in setting the parameters of natural resource use by changing the incentive structures for the poor and the r.on-poor alike. * In all three study areas there was no evidence of widespread vegetative clearings on relatively poor soils. Land use changes were apparent in localized urban peripheries, and within accessible reserve forests in Nigeria. In Sukabumi (West Java), analogous changes were observed in abandoned estates, and in all lowlying reserve forests. On farmlands with steep slopes, agro- forestry, rather than dryland farming appeared to be the preferred income earning strategy in Sukabumi. Overall, changes in land use appear to be rational responses to changing economic opportunities. The Nigerian data, however, requires further field-level study to verify whether commercial farming in the urban peripheries are environmentally sustainable. * 29 - * Infrastructural investments, notably road construction, have played major roles in changing natural resource use. Roads have increased areas under commercial agriculture, enabled entrepreneurs to illegally fell wood from reserve forest lands, and generally increased the mobility of capital and labor. * Despite the widely differing economic and social conditions in Nigeria and Indonesia, in both countries, the major 'software' or social infrastructural achievement has been the spread of literacy. The performance of other poverty-alleviating interventions, such as health care, safe drinking water, credit have not been as satisfactory. The spatial bias of public investments towards urban areas has been more pronounced in Nigeria. * In both countries, the informal sector, rather than marginal farmlands around villages has become the preferred "open access resource" for the poor, and for many of the educated unemployed. * In West Java the growth of circular migration, and spread of non- farm activities have been significant, reflecting greater spatial and occupational mobility of the poor. Despite no observable improvements in income and wealth distribution, poverty in West Java has been alleviated by economic growth and spatial integration of rural areas. Poverty has, however, aggravated environmental problems indirectly in the form of unsatisfactory hygiene (congestion, pollution etc). * The spatial dynamics of poverty has been strongly influenced by the distance of rural communities from cities, availability of transport links, and the growth of production and consumption linkages. In areas of high population density like Sukabumi (West Java), the entire rural hinterland appears to have achieved spatial integration, creating large markets for commodities and labor. In the Nigerian case studies, spatial integration appears to be confined to a five to ten kilometer radius around major towns (areas of high population densities). Land clearing was also observable along major road networks, even in areas of low population density. Vegetative cover changes in physically isolated areas were not observable in either Gombe or Ekiti-Akoko. * In both countries, the growth of population has been mainly confined to existing settlements. In Gombe this feature is noticeable despite the large areal expansion of farming onto the heavy clay soils of river valleys. Validity of the Study for Research and Qterational Use: 5.5. For many developing countries, natural resource endowments are the most significant component of national wealth. Unless usage of these endowments are carefully monitored, there is a danger that poorly designed policies, rapidly increasing populations and poverty could irretrievably - 30 - damage their capital stock. As all physical investments ultimately have a spatial destination, one could, in theory, utilize spatially referenced socio-economic and demographic data sets to: (1) Explain why a certain pattern of change in natural resource stock is occurring, (ii) Identify where unsustainable use of natural stock is likely to be taking place. Consequently, (iii) Given the financial constraints, identify areas that are better suited for agricultural and other economic activities, (iv) Evaluate the costs and benefits of re-structuring economic activities identified by (iii) above. 5.6. In a typical developing country, however, the practical problems of implementing a research agenda discussed above appear formidable. Data on both socio-economic and natural resource variables are usually unavailable at sufficient detailed level (cadastral information or even individual settlements). Collecting these data sets are also unaffordable because of competing demands for limited budgetary resources. What this desk study tried to show was that despite these practical problems, existing sources of data could be combined for a more aggregative, regional-level analysis of natural resource use-poverty-public policy linkages. These studies provide policy makers techniques that are useful in both project and sector work. As and when data sets improve in quality, the analysis would generate more precise information that could serve the needs of infrastructural planners as well. 5.7. In conclusion, many macro-economic policy instruments, such as investments in infrastructure, taxes and subsidies result in stock changes of physical capital in identifiable areas. Once the impacts of economic policy instruments are translated from conventional monetary aggregates to spatial 'stocks' or areal densities (see Beckmann and Puu 1985), it is possible to combine them with demographic data, and study their linkages with natural resource use. By reducing all relevant data sets to spatial statistics, one can integrate economic and natural resource analysis techniques so that future policies and investments remain in harmony with sustainable development objectives. - 31 - Adepoju, Aderanti; "Migration and rural opportunities in Nigeria", Liag and Society, Vol 9, No 1, January-March 1984, pp 29-49 Adepoju, Aderanti; Selected Studies on the Dynamics. Patterns and Consequences of Migration. IV. Medium-sized Towns in Nigeria: Research and Policy Prospects Reports and papers in the social sciences UNESCO (1982) Agunbiade, A.0.; Poverty and natural resource degradation in Nigeria. a case study of Ekiti-Akoko ADP papers prepared for the Environment Department AMEC Inc; Niaeria Processina in the proosed ADPs (1984) APMEPU; Project Completion Renort: Ekili/AkM Aricultural Development Project (1987) APMEPU; Project Completion Report: Gombe Agricultural Development Project (1982) Barbour,K.M.; Plannina for Nigeria A Geograohical Approach, Ibadan University Press (1972) Bates,Robert H; Markets and States in tropical Africa Berkeley: University of California Press (1981) Beckmann, Bjorn; "Public Investment and agrarian transformation in Northern Nigeria" in State. Oil. and Ariculture in Nigeria, Michael Watts ed. (1987) Beckmann, M. and T.Puu; Spatial Economics: Density. Potential and Flow: Amsterdam: North Holland; Studies in Regional Science and Urban Economics (1985) Berry,Sara S.; "Oil and the disappearing peasantry" in Michael Watts ed (1987) Biro Pusat Statistik; Statistical Yearbook 1986, 1976 Brown,Foster, Tom Stone; "Using satellite photography for grassroots development in Amazonia" Cultural Survival Quarterlv, Vol 13(1), pp 35-38 (1989) Chambers, Robert; "Sustainable Rural Livelihoods: A Strategy for People, Environment and Development", IDS Sussex Commissioned Say no 7 (1987) Clough P. and G. Williams; "Decoding Berg: The World Bank in rural Northern Nigeria" in Michael Watts ed. (1987) - 32 - Collier, Paul; "The oil slump and wage rates in Nigerian Agriculture" Paper prepared for the World Bank Collier, Paul; "An analysis of the Nigerian Labor market" Discussion ,aper No DRD 155, Development Research Department (1986) Daly, Herman; Initiating Memorandum on Poverty and Environment (draft) Dalton, George ed. Research in Economic Anthropoloy, Vol 4, New York: Jai Press (1981) Douglass, Mike; "Changing patterns of access to Agricultural land on Java and the outer islands"; Paper prepared for the World Bank (1987) Durojaiye, B.O. and John Aihonsu; "Market integration and seasonal prices of staple foodstuffs A case study of Ogun state, Nigeria" Food.Policy, Nov 1988 pp 375-382 Essang S.M.; "Determinants and impact of rural-urban migration: a case study of selected communities in Western Nigeria" African Rural Emploment Paper N_10, Michigan State University (1974) Faborode,M.O.; "A study in developing country mchanization The size of one state's problem in Nigeria, World Grop, November/December 1984 pp 205-208 Federal Office of Statistics, Nigeria;Annual Abstract of Statistics (1985) Glewve,Paul and van der Gaag; gontronting Poverty ia Develovin Countries, Living Standards Measurement Study Working Paper No.48 (1988) Haggblade S. and Peter Hazell; "Farm/nonfarm linkages in rural sub-Saharan Africa. Empirical evidence and policy implications" AGRAP Economic Discussion Papers No 2 (1987) Hartmann, Joeg; "Landlessness and rural employment in Indonesia" WCARRD Follow-up Programme, tdg4 A Agricitural Oranization 1985 Haskoning Consultants; Environmental profile, Wt Java. Indonesia 1987 Hugo,Graeme J.; foulation mobilit JA WA Jav&, Gadja Mada University Press 1981 - 33 - Health Division, Ondo State; Demographic and Health Survey 1986 Preliminary Report Jagannathan,N.Vijay; Informal Marketsaa DeveloinA Countries, Oxford University Press, New York (1987) Jagannathan,N.Vijay; "Planning new towns: the Durgapur experience", lgnomi nd Political weeki, March 28, 1987 Jagannathan N.Vijay; "Key issues for a studyt of overty-environment linkages in Nigeria", "boMertX-environment linkages. case study of Gombe", *PavertX-Enviromen linkages: Case study of West Java"; Papers prepared for the Environment Department Jamal V. and Weeks,J; "The vanishing rural-urban gap in sub-Saharan Africa" International Labour Review, Vol 127, No3 (1988) Jodha, N.S.; "Common Property Resources and Rural Poor in Dry Regions of India", Ecnaic A P olitical Weekly, Vol XXI No 27,July 5 1986 Kahnert,Friedrich; "Improving urban employment and labor productivity" World Bank Discussion Paper No 10 1987 Kuklinski,Antoni ed; Regional Policies in Nigeria. India and Brazil, Mouton publishers, The Hague (1978) Lipton, Michael; Labr Ind PovertX, World Bank Staff Working Papers Number 616 (1983) Lipton, Michael; Land Assets agn Ral Pover&y , World Bank Staff Working Papers No 744 (1985) Manning, Chris; "Rural Economic change and labor mobility: A case study from West Java", Bulletin of Indonesian Economic tIAes, Vol XXIII, No 3, pp 52-79 Hears, Leon A.; The new rice economr of Indonesia; Gadjah Mada University Press 1981 Mortimore,Michael; "The causes, nature and rate of soil degradation in the northermost States of Nigeria and an assessment of the role of fertilizer in counteracting the process of degradation", Environment Department Working Paper No 17, World Pank (1989) Oludimu,O.; "Focus on Agricultural credit financing in the rural economy of Nigeria - A case study of Oyo region of Nigeria" Nigerian Journal of Financial Management Vol 2 No 2 (1983) pp 25 - 36 - 34 - Patten Richard, Belinda Dapice and Walter Falcon; "An experiment in rural employment creation: The early history of Indonesia's Kabupaten Development" Bulletin of Indonesian Economic Studies, Vol XIX No 3, Dec 1983 Pean, Leslie; Working gaper on the Urban informal sector in the laal, Infrastructure Operations, Sahelian Department World Bank (1989) Remenyi,J.V.; "Issues in smallholder dairying", Bulletin of Indonesian Economic Studies, Vol XXII. No 1, April 1986, pp 57-87 Riddell,kobert; Regional Develoment Policy, New York:St Martin's Press (1985) Rietveld, Piet; "Non-Agricultural activities and income distribution income distribution in rural Java" BulletIn of Indonesian DeveloIent Studies Vol XXII, No 3 1986, pp 106-117 Rietveld, Piet; "Urban development patterns in Indonesia", Bulletin of Indonelian D2eloument Studies Vol XXIV No 1 1988, pp73-94 Rondinelli, Dennis A. and Kenneth Ruddle; Urbanization and Rural Dalomnt, New York: Praeger Publishers (1978) Ross, Paul J.; "Land as a right to membership" in Michael Watts ed. (1987) Siamonds, Norman W.; Farming Systems Research A Review, World Bank Technical Paper Number 43 (1985) Sjahrir; Basic Needs in Indonesia, Institute of Southeast Asia studies, Singapore 1986 Tjokrowinoto,M.; "Small village credit program of the provincial development program in East Java", Paper presented at the UNCRD workshop, Nagoya (1982) Tjondronegoro,SMP; Social _ranigation and 21anned development in rural Java, Oxford University Press, Singapore 1984 Ukvu,U.I.; "Trade, marketing and development in modern Nigeria" Nigerian Journal of Development Studies, Vol 2 No 2 (1982) pp 170-187 Watts, Nichael; Introduction in Michael Watts ed. (1987) World Bank; Rural Develoment World Bank experience 1965-86 (1988) World Bank; Workshop on the Urban Energy Transition and Househtold interfuel substitution (1989)