This section gives an analysis of the characteristics of poor agricultural producers in the context of their role as managers of natural capital assets, analysing the impacts of poverty on the preferences and constraints of farmers related to investment decisions. The analysis includes an assessment of the differences between the investment resources of poor agricultural producers as compared with the non-poor, as well of the differences in attitudes or preferences among poor agricultural producers that might affect investment decisions. Together, these make up an investment "profile" of poor agricultural producers.
In order to begin this profile, it is necessary to describe in greater detail what is meant by the term "poor agricultural producer". Agricultural producers may be owners or tenants. Landless agricultural wage labourers are not considered, even though they are often among the poorest rural households, because they in general do not make decisions regarding the management of the natural resources they work on. Poverty has been defined in many different ways, using traditional indicators based on income or consumption, as well as indicators based on social factors such as health, education, political participation or isolation. For the purposes of this paper, no specific definition is adopted, as much of the discussion is based on a wide range of studies that use varying poverty indicators.
The concept of poverty that underlies the analysis in this paper is based upon an assessment of the asset holdings of the household according to the following categories: (i) natural capital; (ii) financial/manufactured capital; (iii) social capital; and (iv) human capital. These assets are employed by members of the farm household, together with their labour, to generate flows of wealth or income. A lack of assets in one or more of these categories results in an inability to generate an income stream sufficient to meet both consumption needs and the capacity to make investments necessary to attain a sustainable increase in income over time. Here, poverty takes on a somewhat broader definition than is often the case, in that it includes an incapacity to make investments which may improve future welfare, as well as the incapacity to meet a minimum level of welfare in the current period (Reardon and Vosti, 1997). The focus in this paper is on how various manifestations of poverty have an impact on the investment capacity of the household, particularly with respect to their soil resources.
How does the endowment of natural capital among poor farmers compare with that of the relatively wealthier? In this section the prime concern is with analysing the differences in the quality of biophysical environment under which poor and non-poor farmers operate. The implications in terms of investment incentives are whether poor producers have a systematically lower return to investments in natural capital because of the quality of their endowment. Obviously the quantity of land will be an important determinant of this investment capacity as well, and the lack of access to land among poor agricultural producers is well documented. Indeed access to land is often considered to be the most significant determinant of rural poverty. However, in this analysis, it is interesting to expand the discussion to include considerations of land quality. That is, aside from having less land, do poor farmers have poorer quality land, or land located in areas of less agroclimatic production potential and, if so, how does this affect their incentives and capacity to make investments?
Spatial correlation
Analysing the spatial distribution of poverty and environmental endowments leads one into the centre of the debate on the causal links between the two. The analysis is complicated by the fact that human interventions have an impact on land resources and agricultural productivity, so that even if significant correlations are found between poverty and environmental endowments, it is not clear what they arise from: are people poor because the resource base is poor, or is the resource base depleted because poor people have degraded it? The situation is even more ambiguous if migration and the dynamic nature of resource degradation are taken into consideration, e.g. poorer populations may move into areas which have been degraded by wealthier inhabitants who have since moved on.
There have been several studies ranging from the macro to micro level investigating links among natural resource endowments, poverty and environmental degradation. In some cases the studies were focused solely on finding correlations, while others on causality. In some studies strictly exogenous features of natural capital endowments are considered in relation to the incidence of poverty, while in other cases, agricultural productivity which includes factors endogenous to the incidence of poverty have been considered. With this latter type of investigation it is difficult to sort out causal relationships.
Scale and heterogeneity are two other critical issues that must be considered when analysing the spatial correlations among variables. The variations over space in soil quality and rainfall distribution are different, with soil quality generally showing higher variability within a given land area. The spatial distribution of rural poverty can also be quite heterogeneous; in some cases income levels can vary quite substantially within villages, while in other areas income levels may show more significant variation between villages or even larger administrative units such as counties or provinces. Thus the scale at which data are gathered has considerable implications for the type of effect that can be effectively measured.
In the remainder of this section a series of articles on correlations and causal links between poverty and natural capital endowments is reviewed. The features of natural capital of primary interest here are those related to climate and soil, as these have the greatest implications for agricultural production and variability. First, studies focused on exogenous climatic endowments and the distribution of poverty are looked at, followed by studies focused on land and soil resources and poverty on a more micro scale of analysis.8
Broca and Oram (1991) looked at the distribution of poverty across the FAO agro-ecological zones, using caloric consumption (or food expenditure equivalent to meet a minimum consumption standard) as the indicator of poverty. Poverty data were taken from a wide variety of sources including income and consumption surveys and surveys done under various research projects. The scale of poverty data in the study is thus quite variable, from the micro up to the national level.
Two measures of poverty incidence were considered over the agro-ecological zones: (i) the total number of poor people; and (ii) the percentage share of the population in the zone that was poor. Their results indicated that the highest numbers of the poor are located in arid zones and the least numbers in wet zones. The highest share of the poor in the population was found in seasonally dry zones. These zones are somewhat related to agricultural productivity in that the length of growing period for crops varies from year-round for wet zones to under 80 days for highly arid zones. However irrigated area was not accounted for in the agro-ecological zones available at that time, and exogenous factors such as soil type and topography can have a significant impact on agricultural productivity within climate zones, so a direct link among resource endowment, agricultural productivity and poverty could not be made from this study's conclusions.
Sharma et al. (1996) used rates of child malnutrition as a poverty indicator and calculated their distribution over the FAO agro-ecological zones. They found that globally 58.5 percent of malnourished children in developing countries are located in the warm tropics, 26.8 percent in the warm subtropics and 14.7 percent in the cool tropics and subtropics. They did not find a clear link between child malnutrition and agricultural productivity, however; high rates of malnutrition were found in areas of both high and low productivity.
In 1997, GRID-Arendal conducted a study on the relationship between rural poverty and land use potential in West Africa, using geographical information systems (GIS) technology. Poverty was measured using indicators from the Human Development Index9 developed by UNDP, which were overlaid with measures of the "marginality" of lands, including both biophysical (agroclimatic zones and level of land degradation) and socio-economic (population density and accessibility to infrastructure and roads).10 They found a weak correlation between three poverty indicators and degree of aridity, no significant correlations between poverty indicators and land degradation, and a weak negative correlation between child mortality and accessibility.
Kelley and Rao (1995) looked at the distribution of poverty and marginal environments in India using state level data and by using average crop output values to define marginality. They found no relationship between poverty and marginal environments, using both total numbers and share of poor in the population as dependent variables. Fan and Hazell (1999) also found no relationship between poverty and marginal environments in India, using agro-ecological data together with rural poverty data derived from sample surveys of consumer expenditures. These were redistributed over agro-ecological zones through the use of population weights. In this study, both a higher concentration and higher numbers of rural poor were found in rainfed areas of high agricultural potential than in the rainfed low potential areas.
Based on the evidence from these studies, it appears that populations located in areas of arid climates may be more susceptible to poverty, but intervening factors, such as investments in irrigation development, can mitigate this relationship, so that no clear cut correlation between climatic endowments and poverty can be found. This paper, however, focuses on soil endowments and the distribution of poverty and moves on to a review of studies that have focused specifically on this aspect of natural capital endowment and poverty.
In an evaluation of the CGIAR research priority setting, Nelson et al. (FAO, 1997) found that the biophysical characteristics of production settings was not a sufficient criterion on its own for setting research priorities for a poverty alleviation strategy. They developed an indicator of environmental endowments with two values: favoured lands and marginal lands. Marginal lands were defined as those in which the absence of external inputs had resulted in a level at or exceeding the threshold limits to the enhancement of agricultural performance. Favoured agricultural lands were defined as rainfed or irrigated lands which are fertile, well-drained, with even topography and generally under fairly intensive agricultural use.11 They found no correlation between the incidence of rural poverty and the distribution of land quality, leading them to reject the hypothesis that there exists a spatial correlation between poverty and a strictly biophysically defined measure of environmental fragility.
Kirschke, Morgenroth and Franke (1999) tested the causal relationship between poverty and soil degradation using a national level panel data set covering the period from 1960 to 1990 for 73 countries. Using the GIS-based GLASOD data set on human-induced soil degradation as the dependent variable they explored the explanatory power of a set of variables representing socio-economic and technological conditions. As indicators of poverty, they used a wide set of variables, including the average calorie supply per capita, the average percent added value in agricultural production per capita, the density of tractors per hectare and the growth rate of added value in the agricultural sector. None of these variables were found to be significant explanatory variables in the regressions, leading the authors to reject the hypothesis that poverty is a cause of soil degradation.
A higher degree of correlation is found in micro level studies of poverty and soil endowments however, although exceptions exist at this scale of analysis as well. For example, Ravenborg et. al 1999 used household survey data from three watersheds in Honduras and found a strong link between soil quality and poverty. They found that between 25 to 44 percent of the poorest households had their most important food plots on soils classified as the worst quality, e.g. shallow, steeply sloping and highly erosive. Only 8 to 13 percent of the least-poor households reported plots on this quality of land (Ravenborg 1999).
Shepherd and Soule (1998) conducted simulations of the impact of soil fertility endowments on farm income, based on farm model typologies parameterized with household data from West Kenya. Their results indicated that farms with a lower initial endowment of soil fertility would experience declining yields and profits, caused by an inability to replenish nutrients mined over time. In their simulation results, farmers with low and medium natural resource endowments had only 7 to 13% of the farm income of farmers with high resource endowments. However, De Jager et al. (1998) conducted an empirical analysis of 26 households located in three districts of Kenya and found no link between farm income and soil nutrient balances. It is not clear how much these results from a small sample size can be generalized to the population as a whole, and if these contradictory results are due to a greater capacity of low resource endowment farms to replenish nutrients than expected, or the fact that nutrient levels have not declined sufficiently as yet to impact farm incomes.
A different approach to measuring the impact of geographic factors on poverty at the micro level has been taken by Jalan and Ravaillon (2000) . Using a six-year panel data set for China, they regressed consumption growth at the household level on geographic variables, allowing for non-stationary individual effects in the growth rates. They found that there is a strongly significant geographic effect on consumption growth, which they hypothesized was due to "poverty traps". They speculate that these arise as a result of lower returns on farm household investments which, in the presence of borrowing constraints and limits on capital mobility, reduce the growth rate of consumption.
From these studies it can be seen that no unambiguous relationship exists between natural capital endowments and the incidence of poverty, particularly at a macro scale of analysis. While it does seem that "geography matters", the impact of agro-climatic conditions on poverty is mitigated by several factors, principally human investments in productivity enhancing improvements. California and Israel provide two striking examples of how massive investments into irrigation and other infrastructure have resulted in the creation of highly productive agricultural areas, although their agroclimatic endowments could be considered marginal.
Another intervening factor in the spatial correlation between poverty and environmental endowments is the capacity of populations to migrate, among areas of varying quality. There is considerable migration among the poor, but their choice of resettlement location is often constrained. According to Myers and Kent (1995) there is substantial migration among the poor due to an inability to adopt to inhospitable environments. They estimate that there are approximately 25 million such "environmental refugees" in the world, primarily located in the Horn of Africa and the African Sahel, but also found on the Indian subcontinent, in China, Mexico and South America. On the other hand, however, there are indications that in some regions the geographic mobility of the poor is constrained, either by administrative barriers or capital constraints (Ravallion and Jalan, 1999).
Moving to the micro level of analysis however, there is greater evidence of a link between poverty and poor soil quality. Soil quality is highly heterogeneous over space, so micro level studies are necessary to capture the variations among income groups. There are several possible explanations for why a correlation might exist. If land values are tied to land quality, then the poorest quality land may be the most accessible to the poor. Poor quality lands may be the only option available to migrants. In these cases, it is most likely that the poor quality of the land is a determinant of the poverty of the household. In addition, there may be cases where the poverty of the household has lead to a systematic depletion of the soil resources, which would also result in a spatial correlation between poverty and poor soil quality.
A lack of correlation between soil quality and poverty may be evidence of a low level of heterogeneity in soil quality, the failure of land markets to account for land quality, abundant supply of high quality frontier lands, or the impact of past investment efforts on both a public and private scale. Public investments can either directly improve the land quality, or augment the private incentives and capacity to make improvement investment. For example, Fan and Hazell (1997) found a high degree of complementarity between public and private investments in areas of low agricultural productivity, where a unit public investment in infrastructure development resulted in a much higher propensity of the farmer to invest in irrigation and new technologies.
The ambiguity in the macro and micro level research results on the spatial correlation of poverty and soil degradation are due to the lack of data availability at the spatial and temporal scales necessary to measure the relationships in question, as well as the fact that the two processes are jointly determined, which creates problems in sorting out causal directions. Separating out these causal relationships at the micro level in order to understand the impacts of soil quality on poverty and the way in which the dynamic of poverty is manifested in soil quality is crucial for effective policy-making for both the poverty alleviation and improved environmental management. With new GIS-based data sources and improved estimation techniques it is becoming possible to address these questions, although considerable work is still needed in this important area of research.
In this section the importance of natural capital assets as a source of income among poor agricultural producers is examined, and contrasted with that of non-poor producers. Frequently this question has been addressed by contrasting the share of income from farm and non-farm activities among various income groups in the rural sector. No standard definition for non-farm income exists, but generally it is considered to include wage earning from non-agricultural labour, income from non-agricultural related enterprises (e.g. crafts, services) and income from remittances or other transfers. In contrast, farm income includes income derived from crop and livestock production, including agricultural wage labour, and may also include post-harvest processing associated with agricultural production, as well as income from fishery and in some cases forestry production. Clearly the activities in the latter category are highly dependent on natural capital stocks; the income generating activities in the farm sector are biomass based, with natural capital as a major factor of production. Ekbom and Bojo (1999) look at the specific nature of the activities the poor are reliant upon for income generation in the farm sector. They note that the people most likely to be highly dependent on biological resources for income generation include: (i) small-scale farmers; (ii) transhumant pastoralists; and (iii) artisanal fishermen, many of whom are among the rural poor.
Most studies have indicated that farm income constitutes the largest share of rural household income in developing countries, although some exceptions do exist. For 1990, it was estimated that the developing world's share of population mainly dependent on agriculture was 58 percent, which is projected to fall slightly to 47 percent in 2010 (FAOSTAT, cited in IFAD, 2001). Data from farm surveys over the 1970s to 1990s indicated that approximately 70 percent of rural household income in Asia is from farming and approximately 60 percent in Africa and Latin America (Reardon et al., 2000).
How does the share of income from agriculture differ between the poor and the non-poor? The evidence is mixed and again is based on studies ranging from the macro to micro level. FAO (1997) notes that countries which are classified as low income have a much higher share of agriculture in GNP on average and even higher shares of the population in the rural labour force than higher income countries, indicating a high degree of reliance on natural capital assets for their national income generation. Quibria and Srinvasan (1991) conducted a study of seven Asian developing countries and found that the rural poor depended more on agriculture than the rural non-poor.
Reardon et al. (2000) conducted a comprehensive study of the relationship among non-farm income, poverty and inequality. The results of their analysis show quite varied patterns in non-farm income shares and poverty across regions. In general, the evidence indicates a positive relationship between non-farm income share and total household income (or landholding size) in much of Africa, a generally negative relationship in much of Latin America and a very mixed set of results for Asia. They conclude by noting that the farm and non-farm sectors are linked, and that the capacity of the rural poor to enter non-farm sectors depends on their ability to overcome barriers to entry, which is often linked to their farm incomes, as well as the overall performance of the agricultural sector.
The importance of agricultural derived income to the poor, together with the key role of natural capital assets in determining production outcomes, indicates that the poor have even greater incentives to invest in maintaining the services from natural capital assets than do the relatively more wealthy. With less options for alternative income sources, poor agricultural producers have a great motive to protect their capacity for agricultural production, which should create incentives to invest in natural capital stocks, to the extent that such investments are perceived by the farmers to lead to positive production outcomes.
The institutions and social relations that govern the access to natural capital assets influences resource management decisions by impacting decision-makers' expectations, limitations to use and ability to interact and cooperate with other relevant actors. There is considerable variety in the types of institutions governing resource access, depending on the nature of the resource, as well as the society within which it is embedded (Bromley, 1998; Dasgupta 1993). In this paper, the focus is on soil resources, so the institutions of greatest interest are those governing land access, or property regimes. The primary issue of interest here is the degree to which property regimes governing land use are differentiated between rich and poor agricultural producers, and the implications these differences may have for investment behaviour among the poor.
Property regimes can be divided into four categories: (i) private property rights; (ii) common property rights; (iii) state property rights; and (iv) open access regimes (Bromley 1998; Dasgupta 1993 Pearce 1999). The first three categories imply some sort of legal entitlement to land, while open access is the lack of any property right among those who access it. The access of individuals to land within each of the first three categories is quite varied, with some common examples including the following: private properties may be accessed by their owners or by tenants or share-croppers under contractual relations; state property may be accessed by holders of usufruct rights or by members of state-run cooperatives; common properties may be accessed only by members of a local community or by members of outside communities on a periodic basis for restricted uses.
The barriers to entry, or ease with which land can be accessed under any of these regimes are an important determinant of the regime type under which the rural poor will most commonly be found. One important barrier to entry is financial requirements, e.g. the need to purchase an ownership right or pay for tenancy rights. Social barriers to land access exist as well; gender is often a critical determinant of land access as in many rural areas women are denied either de facto or de jure access to land (Jazairy, Alamgir and Panuccio, 1992). Membership in a community group which controls a common property land resource is another social barrier to land access to those outside the community, as are location and lack of mobility.
Common property
Common property regimes (CPRs) are a low barrier-to-entry form of land access to those members of the community which control it. Research from various locations indicates that such access is generally fairly equitable among all group members, including the poor (Jodha, 1986; Quiggin, 1994). Several studies have confirmed that the rural poor are more dependent than the wealthy on land resources accessed through common property (Reddy and Chakravarty 1999; Cavendish 1997; Jodha, 1986; Dasgupta and Maler, 19965). Common property land resources have also been found to be particularly important in providing land access to women (Hopkins, Scherr and Gruhn, 1994; Hecht, Anderson and May, 1988). Areas such as forestlands under community control and commonly managed rangelands fall into this category.
Studies have also indicated that a significant share of the incomes of the poor is obtained from common property resources. For example, among poor rural households in Zimbabwe, as much as forty percent of household income was obtained from non-marketed wild goods collected at local commons (Cavendish, 19987). In contrast, households in the top income quintile derived 25 percent of their income from such goods. Jodha (1986) estimated that between 15 to 25 percent of the income of the rural poor in dryland areas of India came from common property lands, as compared with 1 to 3 percent among the wealthy. The products which are extracted from the commons may vary among rich and poor producers; in Zimbabwe poor households were found to rely more on common lands for the collection of thatching grass and gathered foods, while the wealthier community members with cattle used the lands for grazing (Swallow, 1994).
The management of land resources under common property regimes involves reciprocal externalities, where the actions of anyone who accesses the resource impacts everyone in the group that utilizes the resource (Dasgupta, 1993). In order to attain a sustainable management regime it is necessary for the users to engage in collective action. Otherwise, individual users have incentives to overexploit the resource, as the benefits from doing so are captured by the private individual, while the costs are shared among other present and future users of the resource.
In recent years, a sizeable literature has been generated on the factors that lead to successful collective action in the management of common property resources (Ostrom, 1990; Otsuka; Bebbington and Perrault, 1999; Wade, 1988; Baland and Platteau, 1996). The implications of the presence or absence of social capital on the capacity of a group to take collective action have made up an important part of this literature. Social capital, which is often quite loosely defined, has generally been used to mean the capacity to take collective action arising from the presence of various different sorts of networks based on some form of shared identity and repeated interaction (Bebbington and Perrault, 1999). Networks may occur between families, among community members, or between actors in civil, state and market institutions. An important issue has arisen from this literature: to what extent does social capital complement or substitute for other forms of capital? Of particular interest in the context of this paper is the extent to which social capital may substitute for financial capital in the management of land resources.
There are several ways in which successful collective action in the management of a CPR may serve as a substitute for financial capital, but here only two of the most important which may enter directly into the investment calculations of a land user are focused upon: (i) through unit cost reductions due to increasing returns to scale; and (ii) through the provision of insurance services.
Increasing returns to scale exists when a production technology is more efficient at a higher scale of production. An example of increasing returns to scale is the potential for obtaining lower unit costs of grazing cattle at larger herd sizes due to economies of scale in fencing and herd labour.
The insurance function of CPRs is linked to the physical characteristics of the land resource and its variability over space. For example, in the tropical and subtropical rangelands of the Sahel and East Africa rainfall is highly unevenly distributed over both time and space, so that within a five-mile radius there can be considerable variation in the amount and quality of forage (Swallow and McCarthy 1999; van den Brink, Bromley and Chavas 1995; Baland and Platteau, 1996). Allowing a common access freedom to move grazing animals over a wide area allows producers to pool risks among themselves. In addition, CPRs also have been found to provide insurance through their role as guarantor of subsistence in the case of emergencies (Baland and Platteau, 1996).
In contrast to common property resources where a community or group has a right of exclusion to the resource, open access resources imply that anyone who wants to use the resource is granted the right of inclusion. The decision on accessing an open access resource is based upon the costs of entry as compared with the potential benefits from use. This calculation does not include the costs imposed on other users of the resource impacted by the newcomer; e.g. an externality is generated. However, with open access resources the capacity of the group of users to take collective action is greatly weakened by the inability to exclude others and depletion of the resource is often the result.
Private property
Private property rights have been argued to be the most effective means of providing the proper incentives to individuals to internalize environmentally degrading externalities (Demsetz, 1967). The argument is premised upon the notion that under a private property system the individual bears all the benefits and costs associated with the decisions he or she makes about its use, thereby generating incentives to maximize the value of the land. The problem is that this only holds true under very stringent and unrealistic conditions: (i) that enforcement costs are zero; (ii) that markets are perfect; (iii) that property rights are well-defined; and (iv) that markets are competitive.
There are many examples of actions taken by individuals on private lands that generate negative externalities because one or more of these conditions does not hold. Take, for example the case of soil erosion generated by the land preparation methods a farmer adopts, which result in the siltation of downstream irrigation facilities. To the extent that the farmer has knowledge about the impacts of the erosion on his/her own agricultural production he/she will internalize the costs of engaging in eroding practices into his/her decision-making calculations. However, the costs that accrue to the downstream users will not be considered - no property right is assigned to the eroded soil that enters their irrigation system. Other issues which arise in the efficiency of private property rights regimes are the costs associated with defining and enforcing rights which may be prohibitively high, and distortions that can arise from asymmetric information among agents bargaining over the values associated with the resources (Baland and Platteau, 1996).
An important type of private property regime which offers low barriers-to-entry to the poor is sharecropping or tenancy. This is a very common means of land access for poor producers in Asia and Latin America, although not very common in Africa (Jazairy, Alamgir and Panuccio, 1992). Theoretical results from economic models indicate that an important determinant of the effects of sharecropping on soil productivity is the extent to which soil productivity is reflected in land values. If it is, then practices adopted on-farm to increase soil productivity will result in a greater share of the value of output going to the landowner than the sharecropper.12 There is little empirical evidence on the impact of sharecropping on soil management; however one study from the central plains of the United States did find that crops which do not require heavy working of the soil and that require field drying were more likely to be farmed under cash rental contracts than crops which require tillage and soil manipulation to enhance current profits, which were more likely to be under sharecropping arrangements (Allen and Lueck 1992).
Evolution of property rights regimes
Property rights regimes evolve over time in response to pressures both external or internal to the societies in which they occur. In recent years, rapid changes have resulted in a tremendous pressure for change in property regimes, particularly in developing countries. Changes in technology, increasing population pressure, government sponsored resettlement or privatization schemes, development projects, wars and environmental disasters have all be cited as factors which have resulted in radical changes in property regimes (Dasgupta, 1993;1996 Otsuka et al., 1996; Bromley, 1998; Blaikie and Brookfield, 1987). Two issues arise from these changes which are of particular concern here: the extent to which these changes result in a change of the access regime among the poor, and the impact of the changes in property regimes on the attitudes and capacity of the poor to manage their land resources.
The destabilization of CPRs due to external shocks has in many cases had a negative impact on the capacity of the users to take collective action and manage the resource sustainably. Under external pressures, collective action under a CPR is vulnerable to deterioration into a situation of open access in which there is overdepletion of the resource. This situation in turn has led to calls for privatization of such lands in order to create incentives for better management and reverse environmental degradation. However, in recent years it has increasingly been recognized that in many circumstances creating the means for successful common management of property can result in both improved environmental outcomes as well as positive impacts on the production capacity of the poor. In the move from a CPR to private property the poor may be excluded due to lower capacity to purchase land rights. Even in cases where privatization programmes included distribution of lands to the poorest, the inability to manage these lands productively and maintain household consumption can result in the necessity to sell the land and thus a concentration of land resources into higher income groups (Lipper, 1996).
External forces which result in high population land ratios, low agricultural productivity and increased landlessness create the imperative among the poor to look for low barrier-to-entry possibilities to earning income or food. One potential option is to "squat" on state lands, which are accessible due to the state's incapacity to monitor and enforce its property rights. This may take the form of squatter settlements in forest areas, or temporary incursions into the forest for product collection or shifting cultivation. This type of access is highly vulnerable to becoming an open access property regime and experiencing rapid resource depletion as it is difficult for the users accessing the land to organize themselves into groups that can take actions collectively.
Land reform is another important case in which the evolution of property rights has important implications for the poor in terms of access to land and incentives for soil management.
Access to financial capital among the poor, or the lack of it, is perhaps the most widely researched aspect of asset holding and poverty. For the purposes of this paper, the discussion is focused on access to financial services including credit, insurance and savings. Access to these services is analysed in the context of its impact on the management of natural capital, particularly: (i) the implications for investment constraints; (ii) the impact on risk management strategies; and (iii) the impact on the discount rate, or the intertemporal rate of time preference. All three of these factors have been cited as important determinants of investment patterns in natural resource management.
A considerable amount of theoretical and empirical research has been conducted on the failures of financial markets in the rural areas of developing countries and the implications for the rural poor (Pender, 1996; Holden, Shiferaw and Wik, 1998; Binswanger and Rosenzweig, 1986; Dasgupta and Maler, 1996 1995). One major failure in credit markets is the lack of information on the part of lenders about potential borrowers, and thus the potential for moral hazard and adverse selection. One way of overcoming this problem is for lenders to require collateral for loans, which creates a bias against those poor in asset holdings (Holden and Binswanger, 1998). A second failure occurs because of the high rate of covariance of household incomes within a geographic region, which together with the segmentation of markets due to isolation and difficult transportation leads to disincentives to the provision of credit and insurance services. Would-be lenders or insurers must keep large amounts of cash available to meet their insurance and credit needs, which carries significant opportunity costs. From the borrowers' perspective, the supply of credit would contract at exactly the same time it is most needed, resulting in higher interest rates outside the reach of the poor. Segmentation of markets can also result in the rise of monopoly power among lenders and the ability to charge high interest rates on debt. A third failure occurs because of the high transaction costs associated with providing financial services to the poor, whose size of transaction tends to be quite small. The high ratio of transactions cost to transaction size discourages both lenders and borrowers, and thus significantly lower participation rates in formal sector transactions among the poor (Matin, Hulme and Rutherford, 1999). Finally, barriers to obtaining financial services due to gender or racial characteristics is another failure which results in lower access among the poor.
Faced with these obstacles to entering formal financial markets, the poor instead rely upon a variety of informal mechanisms to meet their needs. These include the use of moneylenders, traders or kinship relations for credit services, and a wide variety of risk management and savings mechanisms (Matin, Hulme and Rutherford, 1999). In recent years there has also been a significant growth in micro-finance institutions such as rotating savings and credit associations (ROSCAs), which are designed to fill the gap in financial services to the poor. Very little research has been conducted on the impact of these new sources of micro-finance on natural resource management patterns among the poor, although as the discussion below indicates, they could potentially have a very significant effect.
An important feature of the utilization of financial services among the poor is the degree to which credit, insurance and savings are substituted for one another, as a result of failures in financial and other markets. Credit and savings are often used for insurance purposes (Matin, Hulme and Rutherford, 1999). Udry (1994) found that credit transaction in Nigeria served as insurance substitutes, with repayment contingent upon production conditions. FAO (2000) presents a theoretical model establishing the link between credit and insurance for what he calls "quasi-credit" transactions, as well as evidence from several empirical studies. This substitution between the uses of financial services means that the household access and mechanisms employed for savings, insurance and credit must be considered as a whole.
In the following sections, the interlinkages between specific features of the services poor agricultural producers adopt for insurance credit and savings purposes and what their implications are for natural resource management patterns are considered. Two things in particular are examined: how effective are these mechanisms for meeting the financial needs of the poor, and what are their implications for investment in natural resources?
According to results derived from the expected utility model, poorer farmers will be more risk averse than the more wealthy, and these differences in risk preferences will be manifested in investment decision-making (Arrow, 1971; von Neuman and Morgenstern, 1944). This result leads one to expect less willingness to undertake high-yielding but risk-bearing investments among poor than non-poor farmers.
However, empirical evidence has indicated that financial constraints and the ability to insure against risks are perhaps even more important in determining investment decisions than the risk preferences of poor farmers (Binswanger, 1990). Considerable evidence exists on the difference between rich and poor in the ability to cope with risk ex-ante or ex-post to investment decisions (see FAO [2000] for a comprehensive survey of these). The mechanisms that are adopted in order to cope with risk, as well as the lack of capacity to insure fully against risk among the poor, have significant impacts on investment incentives, which vary according to situation.
Binswanger (1990) conducted experiments with farmers in rural India in an attempt to measure differences in risk attitudes across varying levels of wealth. He found that the differences in risk attitudes among groups were too small to be of any predictive value in explaining investment patterns or input usage. He concluded that differences in such behaviour could be attributed to differences in the constraints farmers face in obtaining credit or market access rather than their attitudes towards risk. Wik, Holden and Taylor (1998) investigated the cropping decisions of small farmers and found risk preferences had only a minor effect on choice of crop, whereas missing credit and commodity markets were important determinants. Eswaran and Kotwal (1989) demonstrated theoretically that differences in investment behaviour need not arise from differences in risk preferences, but rather due to a difference in the ability to insure against risks, such as access to credit for consumption smoothing.
In a comprehensive survey of empirical and theoretical studies on rural poverty and risk, FAO (2000) provides considerable insight into the limitations the poor experience in coping with risk. He makes the case that the most important negative impact of rural poverty on welfare is through raising the vulnerability of individuals and households to adverse shocks, which they are unable to insure against adequately. He and others make the point that people are much more likely to be able to insure against idiosyncratic risk arising at a household level such as illness or unemployment, but not against covariate risk arising from drought or environmental disasters (FAO, 2000; Udry 1994; Binswanger and, Binswanger and Rosenzweig, 1986).
Several empirical studies have found that the poor are less capable of insuring against shocks to income and consumption than are the wealthy. Jalan and Ravallion (1999) found a strong negative relationship between wealth level and ability to insure in rural China, where 40 percent of income shocks were passed on to current consumption for the poorest households, as compared with 10 percent for the wealthiest. Binswanger and Rosenzweig (1989) also found evidence that poor households are less able to insure against covariate risk than the wealthy, resulting in a less efficient investment portfolio. Udry (19911994) found that agricultural households do manage to insure against individual risks, but were unable to achieve full insurance coverage for covariate type risks experienced at a village or regional level. Rosenzweig and Binswanger (1993) established that wealthier farmers in rural India were able to insure against even covariate shocks, which was not the case among poor households.
One important type of risk that poor farmers face which has considerable implications for their investment decisions is the risk of having insufficient food for consumption, due to production or market failures. Markets for food in the rural areas of many developing countries are thin and isolated, which result in a high degree of covariance between individual and market supply and volatility in markets. Lack of capacity to insure among the poor, together with a high share of staple food items in overall consumption expenditure, results in a strong incentive to produce food crops, even where cash crops may be a more profitable investment for the farm enterprise. (Fafchamps, 1992) For poor farmers then, food consumption concerns enter into the investment decision, whereas wealthier farmers who have better market access and better capacity to insure against market failure can separate their production and consumption decisions, thus allowing them to move into more profitable enterprises ((Fafchamps, de Janvry, Fafchamps and Sadoulet, 1991).
Higher values on risk reduction does not necessarily translate into higher incentives among the poor to make investments in their soil resources. Obviously one important determinant will be the extent to which the investment results in decreased production risk. Again, the work of (Grepperud, 1996; 1997a; 1997b) sheds some light on the relationship between production risk and soil management decisions. In Grepperud (1996), an analysis is made of the impacts of production and price risk on farmers' soil management decisions within a dynamic stochastic production model. He makes an important distinction between risks arising from production variability in contrast to price variability, and notes that in the former case it is critical to know the relationship between input use and soil degradation, and between soil degradation and production variability. These are expected to vary across technologies and production conditions, so detailed knowledge about the specific farming system and natural capital endowments is necessary before any prediction of the impact of risk on soil management incentives can be made. If the risk impacts of inputs on production variability are considered separately, it can be shown that risk-averse farmers will use less of a risk increasing input and more of a risk decreasing input. To the extent that investments in soil represent an input to agricultural production that results in a decrease in production risk then, it can be expected that the greater the risk aversion among farmers, the greater the investment in soil natural capital. Similarly, to the extent that production inputs result in soil depletions, which in turn increase production risk, they will be used less by more risk-averse farmers.
These results are derived from a model in which production decisions are taken separately from consumption decisions, however. Poor agricultural producers are more likely to have linked production and consumption decisions, so that consumption risk enters into the decision-making process as well as production risk, which can result in different implications for investment decision-making. Shiferaw and Holden (1997) argue that the non-separability of production and consumption decisions will make poor farmers less willing to adopt investments in soil natural capital unless they result in immediate production increases as well. They argue that soil investments that result in decreases or even the same levels of output in return for future increases in soil quality will not be adopted, even if they will benefit from such future improvements. Grepperud (1997a) agrees that production technologies which result in both enhanced soil quality and immediate increases in output are likely to be more attractive to poor farmers who are facing consumption risk, but he also notes that investments in soil conservation can be considered a form of insurance, so positive incentives to make conserving investments may also be present among such farmers. The overall impact of consumption risk on production decisions and soil investments is indeterminant, however, in cases where production technologies include inputs with opposing effects on output and soil degradation.
Faced with a lack of insurance services and the inability or highly expensive cost of borrowing at times of crisis, it is frequently the case that households self-insure. One of the most common ways is through precautionary savings - savings in anticipation of future shocks. Savings can be made in a variety of forms, including assets that are used in agricultural production such as draught animals, or even soil resources. A critical requirement of savings for precautionary services, however, is that they be sufficiently liquid to allow for a rapid response to income shocks. Since soil resources are a fairly illiquid investment, precautionary savings in this asset would be expected to be less common than other more liquid assets. Studies have indicated that assets such as grain stocks, jewellery, cash or livestock may function as buffer stocks (Udry, 1994; Chaudhri and Paxson, 1994 ; Rosenzweig and Wolpin, 1993).
The degree and conditions under which productive versus other assets are used as a buffer in times of income shocks is still an unresolved issue, although unsurprisingly it appears that households are likely to liquidate non-productive assets first, and only move on to productive assets if faced with continuing shocks to income. Liquidating productive assets in response to an income shock is likely to result in a decline in future earnings potential, so that the poor often prefer to reduce consumption rather than face this decline (FAO, 2000). Fafchamps, Udry and Czukas (1998) show that households held onto their livestock even at the height of the 1984 Sahelian drought. Murgai (1997) using data from Pakistan showed that poor households were less likely to use productive assets than durable goods to respond to income shocks. However, these findings contrast with the earlier findings of Rosenzweig and Wolpin (1993), who argue that self-insurance among rural households in India is effected through the sale and purchase of livestock.
FAO (2000) fleshes out the arguments on the impact of precautionary savings on investment incentives among the rural poor. He argues that as the poor need their limited liquid wealth to smooth consumption and deal with consumption shocks, they will resist using it to make irreversible, and often lumpy, investments into their production enterprise, even when such an investment could lead to both higher incomes and lower income variance in the long run. FAO (2000) also notes that the poor do respond to investment opportunities by saving more; however income and consumption shocks often occur before they have sufficient time to accumulate enough capital to make the investment.
Another important risk coping mechanism is the diversification of income sources. Diversification of income sources among poor farmers does occur, with different crops and varieties intercropped, or combinations of animal species within herds. The question is what the poor diversify into - e.g. low risk, low productivity investments versus higher risk but higher productivity choices. FAO (2000) argues that the poor have greater incentives to diversify their farm operations through the adoption of new technologies and crops; however barriers exist which prevent the poor from making the investment choices their preferences would lead them to, e.g. the irreversibility and lumpiness of many investment opportunities, as well as food security concerns. Reardon et al. (2000) agree that according to traditional notions of risk aversion, the poor should be more likely to diversify into non-farm income generation than the rich. Again, they cite the barriers to entry (e.g. investment constraints) which often exist to investment opportunities as one reason why greater income diversification among the poor is not observed. They do note, however, that the presence of risk aversion can also create disincentives to the adoption of new and diverse income sources: lack of experience with a new technology or crop creates uncertainty as to the probable outcome of its adoption to the adopter. Ironically, the presence of this type of risk, together with the constraints to adopting risk mitigating measures results in investment decisions that can increase risk, as well as reduce the overall efficiency of investments.
Risk sharing among community or household or other group members is a common insurance mechanism among the rural poor that can take either an implicit or explicit form (FAO, 2000). There is some dispute over the impacts of such sharing arrangements on investment incentives. Platteau and Hayami (1996) argue that risk sharing decreases investment because it taxes success. They argue that this tax is generally higher in African societies than in Asian ones, which is partially attributable to differences in production conditions and population densities. FAO (2000) counters this argument by noting that risk sharing also lowers the cost of failure to an individual investor; the insurance dimension of risk sharing would most likely outweigh the taxing dimension, resulting in a net positive impact on investment incentives.
To conclude, it has been seen that failures in financial markets have more impact on the poor more than wealthy. The poor are less able to insure with financial instruments, and so are more likely to insure by precautionary savings, or risk sharing among local groups. The need to keep precautionary savings in a liquid form in case of problems creates a strong incentive not to invest in irreversible "sunk cost" investments. To the extent that precautionary savings are held in the form of a productive asset the efficiency of production may be negatively impacted. Group insurance schemes may result in lower investment incentives by "taxing success" but most likely have a net positive impact by reducing the costs of failure. Despite a wide array of risk coping mechanisms adopted by the poor, they are less capable than the rich of insuring themselves, which leads to less efficient investment portfolios.
Theoretical results from the neo-classical discounted utility model indicate that the presence of risk together with the lack of capacity to insure adequately against it can increase the rate at which an individual chooses between current and future consumption. (Dasgupta, 1993; Holden and Binswanger, 19998). The presence of a credit constraint, together with risk in income streams decreases an individual's capacity to smooth consumption and may result in a decrease in the elasticity of marginal utility or an increase in the "pure" rate of time preference, both of which will lead to higher discount rates (Holden, Shiferaw and Wik, 1998). The argument is then made that wealthier people, with greater access to credit markets, are more likely to have lower discount rates and thus a higher propensity to make long-term and environmentally beneficial investments. In contrast, the poor are more likely to have high discount rates and a higher propensity to engage in environmentally degrading practices in which natural capital is used for present consumption at the expense of future consumption.
Several authors have opposed this argument, arguing that poverty does not necessarily imply short time horizons or a higher propensity to adopt environmentally degrading practices. Bromley (1998) notes that while discount rates are important in resource management decisions, they are less important than property rights in determining investment incentives. He shows that net returns under an open access property regime produces incentives equivalent to those under a secure property scheme with infinite discount rates. Prakash (1997) argues that it is not short time-horizons among the poor so much as exogenous factors and misguided policy and administrative mechanisms that are primarily responsible for the environmental degradation ascribed to the poor. Ostrom (1990) notes that locally designed and governed resource management institutions can provide resilience towards risks and exogenous shocks, and facilitate sustainable use over time, even among the poor.
Empirical evidence on discount rates by wealth classes is mixed. Part of the problem may lie in the difficulty of measuring the rate at which people trade off future versus current consumption. Pender and Walker (1990) used experimental games with hypothetical questions, but real payoffs to try to get at discount rates. They found that the rate of time preference was inversely related to wealth, and that rates exceeding 100 percent were common among the poorest members of the sample. Holden, Shiferaw and Wik (1998) also used hypothetical questions in surveys conducted in rural areas of Indonesia, Ethiopia and Zambia. The questions were framed by asking respondents how much they would need to be paid today in order to forego a given amount of money to be paid in one year's time. They also collected data on household consumption and production and other relevant characteristics. They found very high discount rates in all three countries and a systematically higher discount rate among the poorest. They also found a higher average discount rate in Indonesia and Zambia where average income levels were lower than the Ethiopia case site.
Cuesta, Carlson and Lutz (1997) conducted an empirical study of farmer's discount rates in Costa Rica, and based on results from this study, as well as 14 other empirical studies, concluded that there is some evidence of higher discount rates among the poorest. Using contingent valuation and asset-choice models to estimate discount rates, they found that 95 percent of the farmers showed real discount rates in an interval from 15.1 percent to 21.9 percent, and rates decreased with income level.
However a different result was obtained by Lumley (1997) from an empirical analysis in the uplands areas of the Philippines. Using data from a survey of 160 households distributed over four villages, Lumley (1997) tested two hypotheses: (i) the rate of interest paid on debt was an appropriate measure of the discount rate; and (ii) the poor were more likely to have higher discount rates. Individual discount rates were measured through the use of a standard bidding rate approach. She rejects both hypotheses, finding a significantly lower discount rate than the amount paid on debt, and that the lowest 10 percent of income earners had consistently lower discount rates than the highest 10 percent of income earners. Surprisingly, in one of the sites a significantly higher discount rate was found among adopters, while in another a significantly lower discount rate was found. Lumley (1997) discusses the differences in land tenure regimes among the villages and how this may have an impact on the incentives to adopt conservation, but no discernible pattern was found.
Lumley (1997) speculates on the impact of inequality in income distribution within the villages on farmer's attitudes and time preferences, having found that the village with the lowest rate of inequality, but also the lowest poverty rate had the highest soil conservation adoption rates. Boyce (1994) elaborates on the impact of inequality in wealth and power on individual time preferences, with specific reference to environmental goods. He argues that inequality leads to higher rates of time preference for environmental goods among both the rich and poor; in the case of the poor for the standard reasons advanced, e.g. imperatives of day-to-day survival preclude making long-term investments. However, in the case of the rich, Boyce (1994) argues that inequality breeds feelings of insecurity among the wealthy based upon a fear of reallocation, and thus motivations to consume more in the present. Additionally, inequality shifts the asset portfolio selection of the wealthy to external, rather than domestic sources, resulting in a need for rapid domestic environmental good depletion to support acquisition of external assets (Boyce, 1994).
To conclude, the evidence presented in this section indicates that the rate at which people trade current consumption for future consumption varies according to a number of factors, including poverty, property rights regimes, inequality in the distribution of assets and others. So although immediate subsistence concerns and cash constraints associated can raise the value of current over future consumption, it is not the only factor and in some cases not the most important factor driving discount rates.
Agricultural producers are constantly faced with decisions on whether to invest in or deplete their soil capital assets through the production decisions they make. In this paper, the incentives and constraints farmers in general, and poor farmers in particular, face when making decisions about investment into their soil resources have been analysed. The analysis indicates that farmers' investment behaviour is a response to a complex mix of incentives, constraints and behavioural attitudes. Poverty has a significant impact on determining what this mix looks like as it frequently imposes greater constraints on the options open to the farmer, and influences the values farmers place on goods and services, which in turn influences the incentives they have to invest.
From the farm level perspective, investing in soil resources may either not yield sufficient returns to justify the investment necessary, or barriers to adoption may prevent such investments even where they could result in an overall increase in farm welfare. The farm level benefits and costs associated with soil quality and its depletion and the barriers to investment which exist are quite heterogeneous across varying agroclimatic and socio-economic conditions.
The analysis in this paper indicates that poverty is an important determinant of both incentives and constraints to farmers in making investments into their natural resource endowments. While heterogeneity also exists among the production and consumption circumstances of poor farmers, there are some systematic ways in which poverty is manifested in natural resource management outcomes, and these are important to consider in environmental as well as economic policy-making.
One important way in which poverty may be expected to influence the incentives of producers to invest in soil resources is by increasing the value to the farmer of an assured food or income source. Poor farmers have less alternatives for generating income and are more reliant on agricultural production for income and consumption than others, therefore maintaining conditions which promote productivity is of greater importance to them. In addition, poverty may generate higher levels of risk aversion among farmers, such that measures which decrease production variability may have higher value to the poor. To the extent that they result in higher productivity and lower variability investments into soil resources may be more attractive to poor farmers than to the relatively more wealthy.
At the same time, poverty is often manifested in greater barriers to investment which prevent poor farmers from taking the investment decisions they would prefer. In terms of soil management, the most important barriers which poverty may give rise to include a poor endowment of soil quality, poorly defined property rights and a lack of access to financial services. Microlevel studies that capture differences in soil quality at the household level indicate that for some case study areas there is a significant correlation between soil quality and poverty. Lower levels of soil capital endowments among the poor may mean they experience higher costs from degradation in terms of production outcomes. However, at the same time, lower levels of soil natural capital imply a lower return to soil investments, and thus less incentive to invest. The access to land among poor agricultural producers varies significantly from that of the more wealthy, with the poor being more likely to obtain access under common property rights regimes, open access or through share-cropping on privately held lands. One response of poor producers to failures in financial markets is the substitution between credit, insurance and savings. Thus, even if credit packages appropriate to soil investments were made available to the poor, the lack of capacity to insure against consumption risk may result in an unwillingness to participate in production credit schemes, or the diversion of funds made available under such programmes to meet consumption shortfalls, resulting in the incapacity to make repayments on the debt.
The implications of these findings are that poor farmers may have greater incentives to invest in soil resources than others, yet face greater barriers in doing so, resulting in an under-investment from both a private and social standpoint. In other cases of course, it may be the case that the farmer does not invest in soil resources because the private benefit is insufficient to warrant it, although from a social perspective the investment may provide a net benefit. In the case of either poor or wealthy farmers, the private returns to soil investments may be less than the social returns, resulting in a suboptimal level of investment from society's perspective. However, poor farmers, unlike the wealthy, may also exhibit suboptimal levels of investment from a private perspective, due to the barriers they encounter in making investments.
Policies designed to address the underinvestment in soil resources must be based upon a sound understanding of the differences between private and social benefits generated by the investment and how these are experienced across heterogeneous production conditions. Recognition that underinvestment may be caused by barriers to investment as well as externalities is crucial. Moving away from a simple analysis of the costs and returns of an investment to a more holistic analysis of the consumption and production environment the producer is operating under is necessary in order to design policies which can promote higher levels of investment. In particular, it is important to note that there are several types of market failures which have significant impacts on the capacity of farmers to make investments, and while these may not seem immediately connected to soil management per se, they will be critical areas of policy making for improved soil management. Examples of such failures highlighted in this paper include insurance and non-farm employment markets, as well as food markets. Here, as in other studies, the analysis indicates the potentially critical role that food safety nets can have not only on the nutritional status of the population but also on investment patterns and natural resource management.
In recent years considerable interest has risen in designing mechanisms for the transfer of payments from society to farmers to provide incentives for higher levels of investment into natural capital assets. In some cases a market mechanism is the proposed solution, as in the case of carbon emissions and sequestration, while in other cases grants such as those made under the Global Environment Facility are the form of payment. In both cases the idea is to provide financial incentives to farmers to take actions which result in the generation of external environmental benefits.
From the farm level perspective, an important distinction needs to be made between programmes which allow farmers to take actions which generate net positive production impacts, as well as generate external benefits, and those which involve payment to farmers for services which have negative or no impacts on production. In the first case, the programme would need to address some barrier to adoption which is preventing the farmer from taking an action in his/her own best interest, while the second case is simply a transaction between parties for services rendered. Each has a different implication for the farmer; the former is a way of enhancing the current livelihood strategy the farmer is engaged in, while the latter is essentially creating a new livelihood source - e.g. the farmer becomes an environmental service provider. From the farmer's perspective, which of these is preferable will depend on the potential incremental return to production which can obtained from taking an action which generates an environmental service, the potential risk and returns which can obtained from becoming an environmental service provider, and how these compare with the risks and returns associated with maintaining current practices.
From the perspective of the purchasers of the environmental service or good, particularly those engaged in market transactions, the criteria for assessing actions or programmes to finance will be quite different: here actions which generate the highest marginal environmental benefit for amount spent will be the most attractive. These actions or situations will not necessarily coincide with those that yield the highest on-farm benefits or, even more stringently, those that will provide the highest on-farm benefits among poor farmers. Thus, the capacity of poor producers to become viable providers of environmental services, particularly in a competitive market setting, is by no means a given. In order to enhance this capacity, it will be necessary to have a clearer idea on where it is most likely that the interests of the poor farmer and the environmental service purchaser coincide.
From the gaps identified in the analysis presented in this paper, it is clear that a great deal more information is needed about the technical relationships among agricultural production systems, soil degradation and yield outcomes, in order to design sustainable technologies and strategies for sustainable agricultural production. One key issue is the extent to which production inputs that increase yields also cause soil degradation, e.g. the extent to which there is a trade-off between short-term yield increases and long-term soil depletion, and what the implications of such depletion are for future production outcomes. Another is the relationship between degradation and input effectiveness, or the degree to which soil natural capital is a complement to other production inputs. These relationships need to be better delineated, including how they vary by soil type and agro-climatic regime. Of course, since this paper refers to the farmer's decision process, his or her perception of these relationships is even more critical than the actual relationships themselves.
Secondly, there is a need for better information on the quality of the soil resources with which poor farmers are working, in order to understand their behaviour and develop technologies that are suitable to their needs. One important question to address then in analysing the investment behaviour of poor farmers with regard to soil resources is the extent to which the current endowment of soil natural capital varies by level of wealth. Studies of a wide geographic scope but on a scale of sufficient detail to determine the degree to which soil quality and poverty are needed in order to determine whether systematic differences in soil quality exist between poor and wealthy agricultural producers, and thus if there are significantly different incentives facing the two groups in making soil investments.
A third major area in which research is needed is analyses of the causal links between poverty and soil degradation processes. The question of how important soil degradation is in terms of household income and the incidence of poverty is still unresolved, as is what the key determinants of this relationship may be. Likewise, how dynamic forces generated by the existence of poverty may be manifested in soil degradation or other natural resource management outcomes is not well documented. A systematic analysis of the various forms of poverty and how these affect natural resource decision-making is lacking, and research in this area would represent an important contribution to understanding and addressing environmental degradation as well as poverty alleviation.
Finally, a newly emerging area of research which is highly policy relevant is the systematic analysis of the potential for poor producers to engage in the provision of environmental services for payments. Although there are widespread assertions that such payments can result in "win-win" outcomes, e.g. both poverty alleviation and improved environmental management, to date there has been very little systematic research on where such circumstances are most likely to exist, and the criteria needed to identify them. In order to do so, at least three relationships need to be defined: (i) the key determinants of the costs of providing an environmental service and how these are distributed across farmers of varying incomes; (ii) the degree of on-farm benefits associated with actions which generate environmental services; and (iii) the distribution of the potential for actions which generate both on-farm and environmental benefits among farmers of varying incomes. With this knowledge in hand, effective mechanisms to promote both environmental benefits and poverty alleviation can be designed.
Allen, D. & Lueck, D. 1992. Farmland leasing in modern agriculture. Choices, 1:30-31.
Arrow, K.J. 1971. Essay in the theory of risk bearing. Chicago, Illinois, USA, Markham Publishing.
Arrow, K.J. & Debreu, G. 1954. Existence of equilibrium for a competitive economy. Econometrica, 22: 265-90.
Baland, J.M. & Platteau, J.P. 1996. Halting degradation of natural resources. Oxford, UK, Clarendon Press, and FAO, Rome.
Barbier, E.B. 1989. Economics, natural-resource scarcity and development: conventional and alternative views. London, Earthscan.
Barbier, E.B. 1998. The economics of environment and development. Cheltenham, UK, Edward Elgar.
Bebbington, A. & Perreault, T. 1999. Social capital, development, and access to resources in Highland Ecuador. Economic Geography, 75(4): 395-418.
Binswanger, H. 1990. Attitudes toward risk: experimental measurement in rural India. American J. Agricultural Economics, 90: 195-407.
Binswanger, H. & Rosenzweig, M.R. 1986. Credit markets, wealth and endowments in rural South India. Report No. 59. Agriculture and Rural Development Department, World Bank, Washington, DC.
Binswanger, H. & Rosenzweig, M.R. 1989. Wealth, weather risk and the composition of agricultural investment. Mimeo. Washington, DC, World Bank.
Bishop, J. & Allen, J. 1989. The on-site costs of soil erosion in Mali. World Bank Environment Working Paper No. 21. Washington DC
Blaikie, P. & Brookfield, H., eds. 1987. Land degradation and society. London and New York, Methuen.
Bojo, J. 1991. Economics and land degradation. Ambio, 20(2): 75-79.
Boyce, J.K. 1994. Inequality as a cause of environmental degradation. Ecological Economics, 11, 169-178.
Broca, S. & Oram, P. 1991. Study on the location of the poor. Paper prepared for the Technical Advisory Committee to the CGIAR IFPRI, Washington, DC
Bromley, D. 1998. Property regimes in economic development: lessons and policy implications. In E. Lutz, ed. Agriculture and the environment: perspectives on sustainable rural development. Washington, DC: World Bank
Carcamo, J.A., Alwang, J. & Norton, G.W. 1994. On-site economic evaluation of soil conservation practices in Honduras. Agricultural Economics, 11: 257-269.
Cavendish, W. 1997. The complexity of the commons: environmental resource demands in rural Zimbabwe. Centre for the Study of African Economies, Oxford: Oxford University.
Chaudhri, S. & Paxson, C. 1994. Consumption smoothing and income seasonality in rural India. Mimeo. Department of Economics: Princeton University.
Clarke, H. 1992. The supply of non-degraded agricultural land. Australian Journal of Agricultural Economics, 36(1): 31-56.
Colacicco, D., Osborn, T. & Alt, K. 1989. Economic damage from soil erosion. J. Soil and Water Conservation, 44(1): 35-49.
Conrad, J.M. & Clark, C.W. 1996. Natural resource economics, Cambridge, MA, USA, Cambridge University Press.
Cuesta, M., Carlson, G. & Lutz, E. 1997. An empirical assessment of farmers' discount rates in Costa Rica. Environment Department, World Bank: Washington, DC.
Dasgupta, P. & Maler, K.G. 1995. Poverty, institutions and the environmental resource base. In J. Behrman and T.N. Srinivisan, eds. Handbook of development economics, Vol. III. p. 2371-463. Amsterdam, Elsevier.
Dasgupta, P. 1993. An inquiry into well-being and destitution. Oxford, UK, Clarendon Press.
Dasgupta, P.S. & Heal, G.M. 1979. Economic theory and exhaustible resources. Cambridge, MA, USA, Cambridge University Press.
Day, J. C., Hughes, D. W. & Butcher, W.R. 1992. Soil water and crop management alternatives in rainfed agriculture in the Sahel: an economic analysis. Agricultural Economics, 7(3/4): 267-287.
De Jager, A., Nandwa, S.M. & Okoth, P.F. 1998. Monitoring nutrient flows and economic performance in African farming systems. Agriculture, Ecosystems and Environment, 71(1-3): 83-94.
de Janvry, A., Fafchamps, M. & Sadoulet, E. 1991. Peasant household behaviour with missing markets: some paradoxes explained. Economic J., 101(409): 1400-1417.
De Leo, G.A. & Levin, S. 1997. The multifaceted aspects of ecosystem integrity. Conservation Ecology.
Demsetz, H. 1967. Toward a theory of property rights. American Economic Review (paper and proceedings), 57.
Dixit, A. & Pindyck, R. 1994. Investment under uncertainty. Princeton, USA, Princeton University Press.
Ekbom, A. and J. Bojö. 1999. Poverty and environment: evidence of links and integration into the country assistance strategy process. Discussion Paper No. 4. Environment Group, World Bank Africa Region, Washington, DC.
El Serafy, S. 1992. The environment as capital. In R. Costanza, ed. Ecological economics: the science and management of sustainability. New York, Columbia University Press.
Erenstein, O. 1999. The economics of soil conservation in developing countries: the case of crop residue mulching. Wageningen, the Netherlands, Wageningen University. (thesis)
Eswaran, M. & Kotwal, A. 1989. Credit as insurance in agrarian economies. J. Development Economics, 31: 37-53.
Fafchamps, M. 1992. Cash crop production, food price volatility, and rural market integration in the Third World. American J. Agricultural Economics, 74(1): 90-99.
Fafchamps, M., Udry, C. & Czukas, K. 1998. Drought and saving in West Africa: are livestock a buffer stock? J. Development Economics, 55(2): 273-306.
Fan, S. & Hazell, P. 1997. Should India invest more in less-favored areas? Environmental and Production Technology Division Discussion Paper No. 25, IFPRI, Washington, DC.
Fan, S. & Hazell, P. 1999. Are returns to public investment lower in less-favored rural areas? An empirical analysis of India. Environmental and Production Technology Division Discussion Paper No. 43, IFPRI, Washington, DC.
FAO. 1978. Report on the Agro-ecological Zones Project. Vol. 1: Methodology and results for Africa. World Soil Resources Report No. 48. Rome.
FAO. 1996. Agro-ecological zoning guidelines. FAO Soils Bulletin No. 73. Rome.
FAO. 1997. Rural poverty and land degradation: what does the available literature suggest for priority setting for the CGIAR? by S.J. Malik. In M. Nelson, R. Dudal, H. Gregersen, N. Jodha, D. Nyamai, J.P. Groenewold, F. Torres F. and A. Kassam. Report of the Study on CGIAR Research Priorities for Marginal Lands. Consultative Group on International Agricultural Research, Technical Advisory Committee Secretariat, Rome.
FAO. 1999a. Erosion-induced loss in soil productivity and its impacts on agricultural production and food security, by M. Stocking & A. Tenberg. In H. Nabhan, A.M. Mashali and A.R. Mermut, eds. Integrated soil management for sustainable agriculture and food security in southern and eastern Africa. Proceedings of the Expert Consultation, Harare, Zimbabwe, 8-12 December 1997. AGL/MISC/23. Rome.
FAO. 1999b. New concepts and approaches to land management in the tropics with emphasis on steeplands, by F. Shaxson. FAO Soils Bulletin No. 75. Rome
FAO. 2000. Rural poverty, risk and development, by M. Fafchamps. FAO Economic and Social Development Paper No. 144. Rome.
FAO/IIASA. 2000. Documentation of the Global Agro-ecological Zones Program. CD-ROM. Available at www.fao.org/ag/agl/agll/gaez/index.htm
Fisher, A. 1981. Resource and environmental economics. Cambridge, MA, USA, Cambridge University Press.
Goetz, R.U. 1997. Diversification in agricultural production: a dynamic model of optimal cropping to manage soil erosion. American J. Agricultural Economics, 79(2): 341-356.
Grepperud, S. 1996. Population pressure and land degradation: rhe case of Ethiopia. J. Environmental Economics and Management, 30(1): 18-33.
Grepperud, S. 1997a. Soil depletion choices under production and price uncertainty. Discussion Paper 186. Oslo, Statistics Norway
Grepperud, S. 1997b. Poverty, land degradation and climatic uncertainty. Oxford Economic Papers, 49: 586-608.
Hecht, S., Anderson, A.B. & May, P. 1988. The subsidy from nature: shifting cultivation, successional palm forests and rural development. Human Organisation, 47.
Hicks, J.R. 1939. Value and capital. London, Oxford University Press.
Holden, S., Shiferaw, B. & Wik, M. 1998. Poverty, credit constraints, and time preferences of relevance for environmental policy.
Holden, S.T. & Binswanger, H.P. 1998. Small-farmer decision-making, market imperfections, and natural resource management in developing countries. In E. Lutz, ed. Agriculture and the environment: perspectives on sustainable rural development. p. 50-70. Washington, DC, World Bank.
Holling, C.S. 1986. The resilience of terrestrial ecosystems: local surprise and global change. In W.C. Clark and R.E. Munn, eds. Sustainable development of the biospshere, p. 292-317. Cambridge, Cambridge University Press.
Hopkins, J., Scherr, S. & Gruhn, P. 1994. Food security and the commons: evidence from Niger. Washington, D.C., International Food Policy Research Institute. Mimeo.
Hwang, S.W., Alwang, J. & Norton, G.W. 1994. Soil conservation practices and farm income in the Dominican Republic. Agricultural Systems, 46: 59-77.
International Fund for Agricultural Development 2001. Status of the Rural Poverty Report 2000/2001. Rome, IFAD.
IFPRI. 2000. Global study reveals new warning signals: degraded agricultural lands threaten world's food production capacity. News Release: 21 May 2000.
Jalan, J. & Ravaillon, M. 1999. Are the poor less well insured? Evidence on vulnerability to income risk in rural China. J. Development Economics, 58: 61-81.
Jalan, J. & Ravaillon, M. 2000. Geographic poverty traps? A micro model of consumption growth in rural China. Unpublished research report. Washington, DC, World Bank.
Jazairy, I., Alamgir, M. & Panuccio, T. 1992. The state of world rural poverty. Rome, IFAD.
Jodha , N.S. 1986. Common property resources and rural poor in dry regions in India. Economic and Political Weekly, 21(27): 1169-1182.
Kelley, T.G. & Rao, P. 1995. Marginal environments and the poor: evidence from India. Economic and Political Weekly, October: 2494-2495.
Kirschke, D., Morgenroth, S. & Franke, C. 1999. How do human-induced factors influence soil erosion in developing countries? Paper presented at the International Workshop on Addressing the Impact of Agricultural Research on Poverty Alleviation, San Jose, Costa Rica.
La France, J. 1992. Do increased commodity prices lead to more or less soil degradation Australian J. Agricultural Economics, 36(1): 57-82.
Lal, R. 1984. Soil erosion from tropical arid lands and its control. Advances in Agronomy, 37: 183-248.
Lal, R. 1987. Effects of soil erosion on crop productivity. CRC Critical Reviews in Plant Sciences, 5(4): 303-67.
Lal, R. 1990. Soil erosion in the tropics: principles and management. New York, McGraw Hill.
Lal, R. 1994. Methods and guidelines for assessing sustainable use of soil and water resources in the tropics. USDA Soil Conservation Service Technical Monograph No. 21, Washington, DC.
Lal, R. & Stewart, B.A. 1995. Managing soils for enhancing and sustaining agricultural production. In R. Lal and B.A. Stewart. Soil management experimental basis for sustainability and environmental quality. Boca Raton, FL, USA, CRC Press.
Lipper, L. 1996. Land allocation in Lap Thach District, Vinh Phu Province, Vietnam: economic and environmental implications. Institute of International Studies Working Paper Series No. 7. Berkeley, CA, USA, University of California at Berkeley.
Lumley S. 1997. The environment and the ethics of discounting: an empirical analysis. Ecological Economics, 20 (1997) 71-82
Lutz, E., Pagiola, S. & Reiche, C. 1994. The costs and benefits of soil conservation: the farmers' viewpoint. The World Bank Research Observer, 9(2): 273-295.
Matin, I., Hulme, D. & Rutherford, S. 1999. Financial aervices for the poor and poorest: deepening understanding to improve provision. Working Paper No. 9, IDPM, University of Manchester.
Moyo, A. 1998. The effect of soil erosion on soil productivity as influenced by tillage: with special reference to clay and organic matter losses. In H.P. Bume, H. Eger, E. Fleishchhauer, A. Hebel, C. Reij and K.G. Steiner, eds. Towards sustainable land use Volume 1 Advances in geoecology 31. Reiskirchen, Germany, Catena Verlag.
Murgai, R. 1997. Savings response to economic shocks - an inquiry into portfolio choice and investment behavior. Working Paper, Department of Agricultural and Resource Economics, University of California, Berkeley, CA, USA.
Myers, N. & Kent, J. 1995. Environmental exodus - an emergent crisis in the global arena. Washington, DC, Climate Institute.
Norgaard, R. & Howarth, R. 1992. Sustainability and discounting the future. In R. Costanza, ed. Ecological economics. New York, Columbia University Press.
Norse, D. & Saigal, R. 1992. National economic costs of soil erosion: the case of Zimbabwe. Paper prepared for the CIDIE Workshop on Environmental and Natural Resource Management in Developing Countries,C. 22-24 January 1992, Washington, DC, World Bank.
Odum, E.P. 1953. Fundamentals of ecology. Philadelphia, USA, Saunders.
Oldeman, R. Hakkeling, R. & Sombroeck, W. 1990. World map of the status of human induced soil degradation: an explanatory note. International Soil Reference and Information Centre, Wageningen, the Netherlands, and UNEP, Nairobi.
Ostrom, E. 1990. Governing the commons: the evolution of institutions for collective action. Cambridge, MA, USA, Cambridge University Press.
Pearce, D.W. & Warford, J.J. 1994. World without end economics, environment, and sustainable development. Washington, DC. World Bank.
Pender, J. L. 1996. Discount rate and credit markets: theory and evidence from rural India. J. Development Economics, 50(2): 257-296.
Pender, J.L. & Walker, T.S. 1990. Experimental measurement of time preference in rural India. Mimeo. Food Research Institute, Stanford University.
Platteau, J. & Hayami, Y. 1996. Resources endowments and agricultural development: Africa vs. Asia. University of Namur and Aoyama Gakuin University, Tokyo, Namur. Paper presented at the IEA Round Table Conference on The Institutional Foundation of Economic Development in East Asia, Tokyo, 16-19 December 1996.
Prakash, S. 1997. Poverty and environment linkages in mountains and uplands. Reflections on the poverty trap. CREED Working Paper Series No. 12. London, IIED. (thesis)
Quibria, M.G. & Srinvasan, T.N. 1991. Rural poverty in Asia: priority issues and policy options. Mimeo. Manila, Asian Development Bank.
Quiggin, J. 1994. Common property, equality and development. World Development, 22 (11): 1613-1625.
Ravallion, M. & Jalan, J. 1999. China's lagging poor areas. American Economic-Review, 89(2): 301-05.
Ravenborg, H.M. 1999. Poverty and soil management - evidence of relationships from three Honduran watersheds. Paper presentation at the workshop on Assessing the Impact of Agricultural Research on Poverty Alleviation, Costa Rica, 14-16 September. Contact lead author: Centre for Development Research Denmark [email protected]
Reardon, T. & Vosti, S. 1997. Poverty-environment links in rural areas of developing countries. In S. Vosti and T. Reardon, eds. Sustainability, growth and poverty alleviation. Baltimore, USA, and London, Johns Hopkins University Press.
Reardon, T., Taylor, J.E., Stamoulis, K., Lanjouw, P. & Balisacan, A. 2000. Effects of nonfarm employment on rural income inequality in developing countries: an investment perspective. J. Agricultural Economics. 51(2).
Reddy, S.R.C. & Chakravarty, S.P. 1999. Forest dependence and income distribution in a subsistence economy: evidence from India. World Development, 27(7): 1141-1149.
Roe, T. & Graham-Tomasi, T. 1986. Yield risk in a dynamic model of the agricultural household. In L. Singh, L. Squire and J. Strauss, eds. Agricultural household models: extension, applications and policy Baltimore, Maryland, USA, Johns Hopkins University Press.
Rosenzweig, M. & Binswanger, H.P. 1993. Wealth, weather risk and the composition and profitability of agricultural investments. The Economic Journal, 103: 56-78.
Rosenzweig, M.R. & Wolpin, K.I. 1993. Credit market constraints, consumption amoothing, and the accumulation of durable production assets in low-income countries: investments in bullocks in India. J. Political Economy, 101(2): 223-244.
Sadoulet, E. & de Janvry, A., 1995. Behavior and welfare under risk. In E. Sadoulet and A. de Janvry, eds. Quantitative development policy analysis. Baltimore, USA, Johns Hopkins University Press.
Sanchez, P.A. & Logan, T.J. 1992. Myths and science about the chemistry and fertility of soils in the tropics. In R. Lal, and P.A. Sanchez, eds. Myths and science of soils of the tropics. p. 35-46. SSSA Spec. Publ. 29. ASA, SSSA, Madison, WI, USA..
Sandmo, A. 1973. On the theory of the competitive firm under price uncertainty. American Economic Review, 61(1): 65-73.
Scherr, S. 1999. Soil degradation: a threat to developing-country food security by 2020? Food, Agriculture and the Environment Discussion Paper 27. Washington, DC, IFPRI.
Sharma, M., Garicia, M., Quershi, A. & Brown, L. 1996. Overcoming malnutrition: is there an ecoregional dimension? Food, Agriculture and the Environment Discussion Paper 10. Washington, DC, IFPRI.
Shepherd, K.D. & Soule, M.J. 1998. Soil fertility management in west Kenya: dynamic simulation of productivity, profitability and sustainability a different resource endowment levels Agriculture, Ecosystems and Environment, 71: 131-145.
Shiferaw, B. & Holden, S. 1997. A farm household analysis of resource use and conservation decisions of smallholders: an application to highland farmers in Ethiopia.
Swallow, B.M. 1994. The role of mobility within the risk management strategies of pastoralists and agropastoralists. SARLs Gatekeeper Series SA 47. London, International Institute of Environment and Development.
Swallow, B.M. & Mc Carthy, N. 1999. Property rights, and livestock development in Africa: issues and project approach. In N. McCarty, B. Swallow, M. Kirk and P. Hazell eds. Property rights, risk, and livestock development in Africa. Washington, DC, International Food Policy Resource Institute.
Taylor, C.R. & Howitt, R. 1993. Aggregate evaluation concepts and models. In G.A. Carlson, D. Zilberman and J. Miranowski, eds. Agricultural and environmental resource economics. Oxford, UK, Oxford University Press.
Udry, C. 1994. Risk and insurance in a rural credit market: an empirical investigation in Northern Nigeria. Review of Economic Studies, 61(3): 495-526.
UNEP. 1991. Status of desertification and implementation of the United Nations Plan of Action to Combat Desertification. Draft Report. Nairobi, United Nations Environment Programme.
UNEP/GRID-Arendal. 1997. Mapping indicators of poverty in West Africa. Environment Information and Assessment Technical Report 8. Report prepared for the Secretariat of the Technical Advisory Committee to the CGIAR, FAO, Rome.
Van den Brink, R., Bromley, D.W. & Chavas, J.P. 1995. The economics of Caine and Abel: agro-pastoral property rights in the Sahel. J. Development Studies, 31(3): 373-399.
von Neumann, J. & Morgenstern, O. 1944. Theory of games and economic behavior. Princeton, NJ, Princeton University Press.
Wade, R. 1988. The management of irrigation systems: how to evoke trust and avoid prisoners' dilemma. World Development, 16(4): 489-500.
Walker, A. 1992. Biodiversity and ecological redundancy. Conservation Biology, 6: 18-23.
Walker, D.J. & Young, D.L. 1986. The effect of technical progress on erosion damage and economic incentives for soil conservation. Land Economics, 62(1).
Wik, M., Holden, S. & Taylor, E. 1998. Risk, market imperfections and peasant adaptation: evidence from northern Zambia. Discussion Paper No. D-28/1998. Department of Economics and Social Sciences, Agricultural University of Norway.
Yadav, S. & Scherr, S. 1995. Land degradation in the developing world: is it a threat for food production to the year 2020? Draft paper for workshop on: Land Degradation in the Developing World: Implications for Food, Agriculture and Environment to the Year 2020, 4-6 April, Annapolis, IFPRI. 190 pp.
8 Several studies have used the agro-ecological zones (AEZ) methodology developed by FAO to explore the correlation between poverty and natural resource endowments. (FAO, 1978; 1996) The agro-ecological zones were designed as a tool for evaluating the suitability of lands for various types of agricultural production. The zones are defined by soil type, rainfall pattern, solar radiation and temperature. In the continental and global AEZ studies, soil data are derived from the FAO digital soil map of the world, which has a scale of 1:5 000 000. Climate variables are based upon interpolation of weather station data; in the global AEZ study a climatic dataset with a resolution of 30 arcminutes (2 500 km2 at the equator) is used and (digital) maps are overlaid with one another in order to derive zones of homogenous agricultural production potential based on their natural endowments, as well as specific crop production requirements (FAO 1978, FAO/IIASA, 2000).
9 Including child mortality, adult female literacy, primary school enrollment and children with stunted growth rates.
10 The indicator of land degradation was taken from the GLASOD database that has a scale of 1:5 000 000. Agroclimatic zones were defined based upon the ratio of mean annual precipitation to potential evapotranspiration from simulations derived at a scale of .05 decimal degrees (6 km at the equator). Poverty data were taken from sample surveys aggregated to stratification level.
11 The zones were based upon data from the FAO digital soil map of the world. Poverty estimates were derived from UNDP data on the share of poverty in the population at the national level applied to subnational population estimates.
12 The idea here is that the landowner is risk averse, with risk aversion decreasing in wealth. Higher soil productivity values that are reflected in land values result in higher wealth to the landowner, and thus less risk aversion. Sharecropping is modelled as a means by which landowners can manage the risk arising from the uncertainty on the actions the sharecropper takes with the land, so a lower risk aversion coefficient on the landowner's part results in a lower willingness to share the output.
