Measuring impacts of HIV/AIDS on rural livelihoods and food security

http://www.fao.org/sd/2003/PE0102a_en.htm

C. Shannon Stokes, Visiting Scientist
Population and Development Service
FAO Gender and Population Division
Professor of Rural Sociology, Pennsylvania State University, USA

Part 1 of 2

I. FOREWORD

The HIV/AIDS epidemic is affecting all spheres of human activity and behaviour. Because most of the hardest-hit countries are still overwhelmingly rural, the epidemic represents an enormous threat to rural development. The implications of HIV/AIDS for the demography of rural populations (age and sex composition of rural households, life expectancy of rural inhabitants, etc.) are well known. However, the epidemic's effects on food and livelihood security of rural residents are still inadequately understood. Part of the problem is that comprehensive tools to measure such effects of the epidemic have not been fully developed.

The main purpose of this paper is to examine general patterns of the impacts of HIV/AIDS on rural livelihood assets and to propose a set of indicators to measure these impacts. A related objective is to identify indicators for evaluating the effectiveness of mitigation efforts. The paper builds on previous work done by FAO on the linkages between the HIV/AIDS epidemic, agriculture, food security and rural development. The paper was developed by C. Shannon Stokes while a Visiting Scientist at FAO (presently Professor of Rural Sociology, Pennsylvania State University), in close collaboration with staff of the FAO Population and Development Service. We hope the paper will stimulate innovative thinking and inspire new interventions to effectively counteract the impacts of HIV/AIDS on rural livelihoods.

Marcela Villarreal

Chief, Population and Development Service
and FAO Focal Point on HIV/AIDS

CONTENTS

I. FOREWORD
II. SUMMARY
III. INTRODUCTION
A. The sustainable livelihoods approach
B. Household-level effects
Human capital effects
Financial capital effects
Natural capital effects
Social capital effects
Physical capital effects
C. Community-level effects
Community-level human capital effects
Community-level financial capital effects
Community-level natural capital effects
Community-level social capital effects
Community-level physical capital effects
IV. POSSIBLE INDICATORS FOR THE EFFECTS OF HIV/AIDS ON RURAL LIVELIHOODS
Table 1. Possible indicators of household- and community-level impact of HIV/AIDS on rural livelihoods assets
A. Methodological considerations
B. Mitigation strategies in the agricultural sector
Labour-saving technology
Knowledge Preservation and Transmission
Rural institutions/capacity building
Gender equality
Improving nutrition
Social and economic safety nets
V. IMPROVING IMPLEMENTATION: MONITORING AND EVALUATING IMPACTS OF MITIGATION STRATEGIES
Table 2. Selected indicators and recommended methodologies for evaluating and monitoring mitigation strategies
VI. CONCLUSIONS
VII. References

II. SUMMARY

The HIV/AIDS epidemic has been demonstrated to have far-reaching effects across all sectors of society, but particularly on labour-intensive sectors such as agriculture. FAO estimates that seven million agricultural workers have died from AIDS since 1985 and that another 16 million of the agricultural labour force in sub-Saharan Africa could die by 2020. Thus, the epidemic currently has and is expected to have devastating impacts on millions of rural, farm households over the next two decades.

Most ameliorative efforts have been concentrated in the health sector and have been directed toward the prevention of infection. Comparatively few efforts have been directed toward mitigating the impacts of the epidemic on those households and communities already suffering from the effects of the pandemic. This paper focuses on measuring the impacts of HIV/AIDS on rural livelihoods and food security as a necessary step in the development, monitoring and evaluation of mitigation efforts.

The analysis uses the sustainable livelihoods framework in the development of indicators to measure the impacts of HIV/AIDS on rural households and food security. This framework posits that all rural households possess five sets of livelihood assets, capabilities and activities through which they seek to earn their living. Each of the five capital assets - human, financial, natural, social and physical capital - is demonstrated to be impacted by the epidemic. Indicators to identify and measure the effects of the epidemic on these capital assets are important to the development of more effective mitigation strategies and programmes. Human capital impacts are identified as central to any effort to measure effects of the epidemic because declines in human capital reverberate throughout the other capital assets. Rural communities are also severely affected by the epidemic and indicators of community-level impacts are noted.

Given the range of impacts of the epidemic on individuals, households and communities, a number of alternative indicators and research methodologies are required to measure the impacts of the epidemic and to monitor and evaluate the effectiveness of mitigation strategies. Methodological problems of sorting out the effects of HIV/AIDS from other deleterious effects are described and alternative approaches to their resolution are suggested. Control of alternative explanations and identification of households that have experienced an AIDS-related death or illness are particularly problematic. Combinations of qualitative and quantitative research methodologies appear to offer good possibilities for addressing these problems.Six sets of mitigation strategies were suggested by the Expert Meeting at FAO in December 2001. Each of these strategies consists of a set of programmatic recommendations to alleviate problems created by the epidemic. The labour shortage caused by the illness and death of household members is one of the most pervasive and well-documented losses to households' human capital assets. Thus, use of labour-saving technologies represents an important mitigation strategy.

Recommendations include the use of technologies that would save both agricultural and household labour. They include small farm mechanization, low-input agriculture, lighter ploughs and tools that can be used by older children, women and the elderly, improved seed varieties that demand less labour for weeding, intercropping and zero or minimum tillage, as well as access to potable water and fuel-efficient stoves that can free women for more economically productive activities.

A baseline study of the use of these practices would be important in establishing how many of the recommended practices are currently being used. If resources permit a panel of households to be followed over time, the adoption of selected labour-saving practices can be monitored to see if recommended practices are being adopted and the extent to which the new technologies have alleviated the labour shortage at the household level. Important issues to be addressed by such a study include: Have adopting households increased or maintained the area cultivated, their cropping patterns, yields, and continued to farm more intensively than households that did not adopt the recommended technologies? Evidence from such a study would permit strong inferences about the effectiveness of the strategy on mitigating the labour constraint.

In addition to labour-saving technologies, the Expert Meeting at FAO in December 2001 recognized the loss of agricultural knowledge, practices and skills as an additional human capital loss created by the epidemic. Recommendations to preserve knowledge and transmit it across sexes and generations were presented. Extension education programmes, as well as informal community organizations, need to be reoriented to meet the informational needs of adult-loss households. Orphan- and female-headed households, as well as widows and widowers, need information to be able to maintain agricultural production. The abilities of adult-loss household members to draw up cropping plans, maintain animal husbandry practices, store grain, market agricultural production and to be knowledgeable about gender-specific production practices need to be monitored and evaluated if effective strategies to meet the informational needs of these households are to be designed and implemented.

Mitigation strategies are not only needed to assist individuals and households in coping with the epidemic, but community organizations and institutions also require assistance. Strengthening rural institutions, promoting gender equality, improving nutrition and providing social and economic safety nets were additional strategies suggested. Because the majority of assistance provided to HIV/AIDS-affected individuals and households comes from the extended family, neighbours and local informal community institutions, it is important that these institutions be monitored and evaluated. Some communities have been extremely responsive to the epidemic and institutions have been strengthened and created to deal with a variety of problems created by increased morbidity and mortality. Informal institutions centering on traditional labour-sharing practices and communal farming, self-help efforts to provide home care for the sick and dying, child care, apprenticeship training for orphans and educational and nutritional assistance to affected households are among the strategies recommended. Monitoring and evaluating the effectiveness of these activities is important if external agencies and organizations are to assist in strengthening these traditional institutions. The recommendation is made that evaluation of mitigation strategies focus on only a limited number of livelihood assets designed to be affected. Selection of a small number of livelihood assets focuses the monitoring and evaluation process, leads more directly to the selection of key indicators to measure their impacts and suggests a choice of methodologies to use in examining their effectiveness.

III. INTRODUCTION

The HIV/AIDS epidemic has far-reaching impacts across all sectors of society. In addition to the well-recognized impacts on the health sector, in the hardest hit countries the pandemic has had devastating consequences for the economy, public welfare, education, government, the labour force, and especially on labour-intensive sectors such as agriculture. Seven million agricultural workers are estimated to have died from AIDS since 1985 and FAO (2002) estimates that another 16 million of the agricultural labour force in sub-Saharan Africa could die by 2020. These macro-level impacts are paralleled by equally negative consequences for rural, agricultural households by increasing both livelihood insecurity and poverty. In addition, they are reversing decades of hard-won improvements in households' livelihood assets, those capabilities, resources and activities through which rural households seek to earn a living.

The purpose of this paper is to examine the impacts of HIV/AIDS on rural livelihood assets and to identify possible indicators of their impacts that are relevant to household food security. A related objective is to identify potential measures for evaluating the impacts of mitigation strategies in the agricultural sector on HIV/AIDS-induced food insecurity. Accurately identifying and measuring the impacts of HIV/AIDS on rural households is essential to the development of effective mitigation strategies. Moreover, it is important to monitor and evaluate the impacts of mitigation strategies if their beneficial effects are to be maximized. While the review seeks to identify general patterns for the impacts of HIV/AIDS on rural livelihoods, specific analyses of local contexts are required to apply them in any given situation. Thus, while the loss of household labour through the death of a young adult can lead to a decline in household production in one context, in another situation in which land and labour are more plentiful, hired labour or crop sharing arrangements can maintain agricultural production and insure food security.

The paper takes a sustainable livelihoods approach to identification of the impacts of HIV/AIDS. The framework views even the poorest households as possessing assets that allow them to adjust to shocks (Carney 1998, DFID/FAO 2000). Indeed, households that face food insecurity on a regular basis have developed a series of coping strategies to deal with this problem. Consequently, families and households are viewed as active participants in dealing with the issue, rather than passive recipients of forces beyond their control. HIV/AIDS represents an extreme source of livelihood and food insecurity shock that requires multiple adjustments on the part of farm households to this threat to their survival. Indeed, some have suggested that the effects of HIV/AIDS are so severe and pervasive that households are not able to cope with this trauma in the same manner as with other shocks (Rugalema 2000).

Perhaps the most immediate impact on household assets is the shortage of labour experienced by households in which one or more members suffer from HIV/AIDS. Not only does the household lose the productive labour of the afflicted member(s), it also loses significant labour of other household members whose time is absorbed in caring for the sick and dying member(s). Additional productive labour is lost during funerals and traditional periods of mourning. The difficulty in developing indicators and measuring the effects of HIV/AIDS on household food security stems from the fact that many of the impacts and coping mechanisms employed in response to AIDS are similar to those stimulated by other deleterious events. Nonetheless, by identifying impacts of HIV/AIDS on livelihood assets, we can begin to identify some of the proximate mechanisms through which food security is threatened.

A. THE SUSTAINABLE LIVELIHOODS APPROACH

The sustainable livelihood approach is well-documented elsewhere (Carney 1998, DFID/FAO 2000) and only a brief description is included here. The approach attempts to link the micro- and macro-level contexts in which households seek their livelihoods. The approach attempts to be people-centred, holistic and dynamic. It builds on the strengths that all households are viewed as possessing. Households are seen to possess five sets of livelihood assets essential to their livelihood strategies: human capital, natural capital, financial capital, social capital and physical capital. Utilizing these assets, households adjust to their physical, social, economic and political environments through a set of livelihood strategies designed to strengthen their well-being. The contexts in which households operate involve a number of threats that render them vulnerable to negative livelihood outcomes. These threats can include periodic droughts, floods, pest infestations, crop and livestock shocks, economic shocks, conflict and civil unrest, as well as the illness and death of household members. Households are viewed as being sustainable if they can adjust to threats without compromising their future ability to survive shocks to their livelihoods.

Source: DFID/FAO 2000

HIV/AIDS represents a potentially devastating shock to farm household survival. The illness or death of one or more household members can affect each of the livelihood assets resulting in a reduction in the ability of the household to adjust to future shocks. In countries or regions within countries where the prevalence of HIV/AIDS is high, as in many nations in Eastern and Southern Africa, the epidemic can affect not only the ability of households to cope, but entire communities and regions may find their capacities taxed beyond their ability to respond effectively.

The list of livelihood assets hypothesized to be influenced by HIV/AIDS is not exhaustive, nor is each item necessarily mutually exclusive, rather the list represents an attempt to categorize the impacts on households under one or more of the five asset rubrics commonly included in the livelihoods paradigm. By examining the possible asset classes through which households and communities are affected and respond to the epidemic, we attempt to identify possible indicators of its effects. Household-level effects are examined first, followed by a discussion of community-level effects. Possible indicators of household- and community-level effects on each asset class are developed and methodological considerations in measuring impacts of the epidemic are noted. The final section of the paper examines selected mitigation strategies in the agricultural sector and discusses possible means for measuring their effects.

B. HOUSEHOLD-LEVEL EFFECTS

Human capital effects

The loss of adult on- and off-farm labour is one of the most widely discussed effects of the HIV/AIDS epidemic (Topouzis and du Guerny 1999). The loss of experienced agricultural workers affects both individual households and communities, resulting in labour shortages and declines in productivity both on and off the farm. Declining productivity, in turn, leads to declines in household income through both decreases in the household's own production and through declines in off-farm income and remittances. An increase in household expenditures on medical care results in a decline in savings and the loss of assets through the sale of both productive and non-productive assets. Thus, the loss of human capital leads directly to declines in the financial capital of the household. For food insecure households or those slightly above this threshold, the loss of labour, income and increased expenditures for medical care can push them further into poverty and food insecurity. Among the human capital effects posited to be influenced by the epidemic are the following:

• Illness and/or death of one or more household members
• Change in the size and composition of households
• Children orphaned by the epidemic
• Addition of adult relative to assist with farm production, housework and/or child care
• Temporary migration for wage work
• Change in the household dependency ratio
• Withdrawal of children from school to work on or off the farm for wages
• Intra-household reallocation of labour
• Decrease in area cultivated (increased fallow)
• Decline in crop variety
• Change in cropping patterns and/or animal production to less labour-intensive practices
• Declining yelds
• Lengthening of the working day
• Sex work on a casual or commercial basis
• Loss of agricultural knowledge, practices and skills
• Change in access to human resources

Although this list is not exhaustive, it does represent the range of household human capital assets that are potentially influenced by HIV/AIDS. Some of the asset changes are reversible (temporary migration for wage work) and can actually increase the household's ability to adjust to shocks. Others are irreversible (death of a member) or become more difficult to ameliorate with the passage of time (withdrawal of children from school) and limit the household's future ability to respond to shocks. These losses in human capital reverberate through a household's other livelihood assets, ultimately affecting their livelihood strategies and increasing the likelihood of poverty and food insecurity.

Financial capital effects

The loss of human capital leads directly to a loss of financial capital. Incomes obviously decline as HIV infections and AIDS deaths are disproportionately concentrated in the most productive age groups (15-49). Income thus declines from both farm and off-farm sources, further rendering the household vulnerable to food insecurity. As productive assets are sold off (see Physical Capital below), the household's future livelihood is jeopardized. Topouzis and du Guerny (1999) note that households respond initially by disposal of insurance assets that are reversible, including liquidating savings, seeking remittances from the extended family and borrowing from informal or formal sources of credit. If necessary, the sale or disposal of productive assets typically follows use of these sources of support. Among the financial capital effects thought to be influenced by the epidemic are:

• Reductions in income from farm and off-farm sources
• Liquidation of savings accounts
• Seeking remittances from family
• Change in degree of reliance on off-farm income among male, orphan and female-headed households
• Change in wage earning among female-headed households
• Change in income-generating activities among female-headed households (Topouzis 2000)
• Sale of stores of value (jewellery, household goods)
• Borrowing from informal sector (relatives, friends, neighbours, rural coops, rotating and savings club associations) (Mutangadura et al. 1999)
• Borrowing from rural traders or money lenders (often at exorbitant interest rates)
• Pledging of future crops
• Exhaustion of credit resources
• Sale of livestock

It should be noted that the sale of livestock appears under several asset rubrics and illustrates the difficulty in classifying some effects under only one category. While livestock are generally thought of as part of natural capital, they also operate as a store of wealth in many societies in which financial markets are underdeveloped. Moreover, by providing animal traction power, they also operate much like physical capital assets. Disposal of draught livestock directly affects the household's productive activities and increases the risk of food insecurity.

Natural capital effects

The loss of human and financial capital can have important effects on a household's use and preservation of its natural capital. Access and ownership of land is fundamental to the rural social and economic structure. Indeed, the rural socio-economic structure is often described by the relationship of various segments of the population in their relation to land: landless labourers, share croppers, tenant farmers, landowners and so forth. Farm households are understandably reluctant to sell land because this is the primary natural capital that they control. The distress sale of land is typically the final act of a household on the verge of destitution. However, before this point is reached, there can be serious deterioration in the natural capital of households as the declines in their human and financial capital limits their ability to invest in maintaining and improving their land base.

Natural capital assets that can decline in the presence of HIV/AIDS include:

• Reductions in soil fertility
• Declines in on-farm conservation and/or irrigation practices
• Decreased biodiversity due to asset stripping, selling of firewood, increased harvesting of wild food, game, etc.
• Decline in genetic resources
• Fallow land returning to bush
• Decline in quality of permanent crops
• Renting or leasing out portions of the household's landholdings
• Appropriation of land by relatives (taken from widows, orphans)
• Sale of livestock
• Sale of land

Each of the declines in natural capital reduces the household's ability to adjust to future shocks and the distress sale of land can render the household non-sustainable.

Social capital effects

The illness and death of household members can disrupt a household's links to their extended family and the larger community. In areas where cultural practices limit women's participation in formal organizations outside the home, the death of a male breadwinner can seriously impair a household's ability to access community resources or even receive family support. Indeed, in some cultures upon the death of their husbands women are deprived of their access to land and their husband's resources by his extended family. Nonetheless, studies indicate that households affected by HIV/AIDS draw their support primarily from family, neighbours, community institutions and informal organizations (Mutangadura et al. 1999). Thus, the social capital of households operating through their relationships with extended kin and the community is critical to their ability to recover from the illness and/or death of a household member due to HIV/AIDS. The social capital linkages that can be affected by HIV/AIDS include:

• Relationships with extended family members
• Linkages to formal and informal community organizations (social support groups)
• Community labour sharing for agricultural production, housework and/or child care
• Extended family and/or community willingness to foster orphaned children
• Community willingness to support educational and nutritional needs of orphaned children (school fees, uniforms, supplemental feeding, etc.)

A community's level of social capital can have major impacts on mitigating the effects of HIV/AIDS. Communities with high levels of social capital can provide affected households with a variety of social support activities that permits families to adjust to the illness or loss of members. Conversely, communities with low levels of social trust and solidarity can leave households and families to fend for themselves or even to isolate and ostracize those households afflicted with HIV/AIDS.

Physical capital effects

Households' physical capital refers to those tangible assets and producer goods other than their natural capital which includes their land, biodiversity and water resources. It includes their housing, household goods, furniture, tools and equipment, as well as livestock. Again, households attempt to conserve their productive resources in distress situations for as long as possible. Thus, once savings and credit resources have been exhausted and liquid assets have been disposed of, households resort to selling of other assets. These assets include:

• Housing (condition may deteriorate)
• Livestock (used for animal traction)
• Household goods
• Equipment
• Tools
• Sewing machines
• Bicycles and transport facilities
• Radios

The distress sale of physical capital leaves households in a precarious position in terms of their ability to adapt to future shocks. Moreover, with the disposal of physical assets and equipment needed for agricultural production, households' ability to generate income in the short term is also compromised. Coupled with the disposal of natural capital, particularly land, households that reach this state are frequently dissolved.

C. COMMUNITY-LEVEL EFFECTS

The effects of the HIV/AIDS epidemic can be followed at several levels of analysis. This paper considers two levels: the household and the community. While the most immediately affected units are the individuals, families and households afflicted by the epidemic, these effects combine to exert impacts on communities, regions and nations. The community-level effects are in some cases an aggregation of household-level impacts (e.g. a shortage of experienced labour). In other cases, the household effects can combine to alter the composition and structure of communities' populations (e.g. increased ageing, changing household composition) and influence the price of labour, land and/or credit. The nature of the community-level effects will depend upon how prevalent the epidemic is in the community and upon the local social, economic, cultural and agro-ecological setting.

Community-level human capital effects

Human capital resources form the base from which households and communities are constructed. The severe impacts that the epidemic has for households are paralleled at the community level in the hardest hit nations and regions. Some of the community-level human capital assets that can be affected include:

• Increase in prevalence of illness and loss of significant proportions of the population
• Change in the nutritional status of children
• Shortage of experienced labour force in all sectors of the economy (increase in the price of labour)
• Increase in the number of female-, elderly- and child-headed households
• Increase in the number and rate of orphaned children
• Increase in fosterage rate of children
• Change in out-migration of young adults
• Change in age-sex composition
• Reduction in school enrolment
• Increase in school dropout rate
• Loss of knowledgeable and experienced community leaders

Community-level financial capital effects

Depending upon the prevalence of HIV/AIDS in a region or community, the aggregate impacts on community financial capital could vary from modest effects to significant changes in the aggregate savings rate, credit markets, to reduced expenditures in the local community.

• Decrease in aggregate community income
• Reduction in expenditures in community businesses
• Reduction in aggregate community savings
• Increase in demand for loans and credit from formal and informal sources
• Decrease in demand for productive credit
• Increase in price of credit
• Increase in default rate in credit markets
• Increased spending on traditional and modern health care

Community-level natural capital effects

The epidemic may influence the natural capital of a community through reductions in the available human resources (labour and knowledge) to invest in preservation and conservation. Moreover, the presence and frequency of widespread illness and death may make individuals and communities reluctant to invest in conservation and preservation of natural capital such as biodiversity, community water and land resources that require long investment periods.

• Reduction in quality of community land resources
• Decline in conservation of land and water resources
• Decrease in biodiversity
• Increase in fallow land
• Change in land use patterns
• Change in the land market
• Environmental deterioration, decline in maintenance of infrastructure (e.g. terraces)

Community-level social capital effects

As noted earlier, communities can have a major impact on mitigating the effects of HIV/AIDS on households and families. The strength of social ties, social trust and relationships within a community, the sense of collective responsibility and common outlook, all influence a community's willingness and ability to organize and support affected households. In view of the epidemic, some communities have provided a variety of support and mitigation activities. Some have organized community-based child care, including cooperative day care and nutrition centres to permit women to work outside the home. Others have provided nutritional and educational support to orphans, home care and visitation programmes for orphans and HIV/AIDS patients, apprenticeship projects in marketable skills for orphaned adolescents, labour sharing arrangements and credit schemes for funeral benefits, among other activities (Mutangadura et al. 1999, Topouzis and du Guerny 1999). With increasing prevalence of HIV/AIDS in a community, the question arises as to how long even those communities with substantial social capital can continue to offer such support to affected individuals, families and households. The scale of the epidemic in some regions is such that even the strongest-knit communities may not have the human resources to continue such programmes.

Among the social capital effects within a community are:

• Change/disruption of kinship and extended family ties in the community
• Change/disruption of formal and informal organizations in the community
• Changes/disruption in labour-sharing arrangements
• Increase in demand for community care for sick and dying members
• Change in demand for apprenticeship training for orphaned adolescents
• Increase in demand on community social support and self-help organizations
• Increase in demand for child fosterage within the community
• Increase in demand for child care within the community
• Increase in social inequality (e.g. land acquisition, wealth concentration)
• Community disintegration (breakdown in socialization, cultural transmission)

Community-level physical capital effects

Similar to households' physical capital, communities must continue to maintain and invest in their infrastructure. Community schools, water supplies and other assets require investment and maintenance if they are to serve the needs of residents. While such facilities vary greatly from community to community, with reduced human, financial, natural and social capital, communities may have difficulty maintaining and preserving their limited physical capital.

Possible effects include declines in the condition of community infrastructure:

• Community wells
• Irrigation facilities
• Roads
• Drainage ditches
• Terraces
• School buildings
• Public facilities

IV. POSSIBLE INDICATORS FOR THE EFFECTS OF HIV/AIDS ON RURAL LIVELIHOODS

Table 1 lists each livelihood asset class at both the household and community levels and suggests possible indicators of the impacts of HIV/AIDS. These indicators should be viewed as a first approximation and suggestive of the types of effects the epidemic can have on rural livelihoods. The list is not exhaustive and some effects clearly bridge asset classes. Moreover, these effects can vary greatly depending on the duration, intensity and prevalence of the epidemic in the community and on the local institutional context in which they occur. Application of these indicators should be made only in the context of detailed information about the local situation in which data are to be collected.

Although it is difficult to prioritize the indicators in their order of importance, several considerations enter into the determination as to which indicators should be selected for use in any given situation. First, reducing poverty and insuring food security are both part of the United Nations' Millennium Development Goals. Because preserving and enhancing human capital is so essential to reducing poverty, improving sustainable rural livelihoods and insuring food security, indicators of human capital are essential to virtually every effort to examine and mitigate the impacts of HIV/AIDS. Thus, we recommend that indicators of human capital effects be central to any measurement of the epidemic's impacts. Secondly, the determination must be made as to whether the focus of the study or intervention is to be the household or community (or both). Clearly, different strategies are required if a household survey is to be undertaken or a community-level qualitative investigation is envisioned. Third, the stage of the epidemic and the prevalence rate in an affected community should be used to select indicators. A communities with a comparatively long history of the epidemic and a high prevalence rate can be expected to respond differently than one that is in the initial stage with a low prevalence rate (e.g. below five percent). For example, one would not expect large-scale community-level social capital effects in the latter community, while communities with higher prevalence rates (e.g. 15 percent or more) might well be experiencing marked changes in their social, human and financial capital. Finally, the level of awareness and perceived seriousness of the epidemic by a community is yet another criterion in selecting indicators. Communities that have experienced the loss of significant numbers of their population to HIV/AIDS and that view the epidemic as threatening are more likely to show community-level effects than those places with little awareness, and that are unable or unwilling to discuss the epidemic. Secondary data coupled with participatory rural appraisal (PRA) techniques should be helpful in identifying the prevalence rate, the stage of the epidemic in a given community and the general level of awareness and perception of the threat posed by HIV/AIDS.

In some cases the putative direction of the effect is noted. For example, at the household level, a labour shortage created by the illness or death of a household member is hypothesized to result in a decrease in the area cultivated and an increase in fallow land. In other cases, the direction of the hypothesized effect is indeterminant. The impact of an effect can depend upon local conditions. For example, at the community level, economic theory would predict that a labour shortage can be expected to change the supply of labour and thus exert pressure for an increase in the wage rate for unskilled labour. However, if the epidemic is prevalent and of long duration in the community and significant proportions of the households have changed their market orientation toward subsistence, the financial capital may no longer be available to hire landless labourers, thus depressing the unskilled labour market (Topouzis 2000). Such effects are also likely to vary depending upon the level of inequality and the distribution of wealth and access to resources in a community. Knowledge of the local situation is essential to developing hypotheses about the likely direction and magnitude of an effect.

Table 1. Possible indicators of household- and community-level impacts of HIV/AIDS on rural livelihoods assets

To illustrate how the indicators can be used, it may be useful to consider a hypothetical case of an HIV/AIDS-affected household in a community with a mature epidemic, e.g. one in which 15 to 20 percent of adults are HIV-positive and sufficient numbers have died such that traditional mourning periods have been shortened in order to reduce further losses to agricultural production. The example drawn does not refer to any particular household, community or nation and is presented solely for heuristic purposes in illustrating how the indicators may be used. By focusing attention on the impacts of the epidemic on the five livelihood assets, the sustainable livelihoods paradigm provides a delimited area in which to examine possible impacts and develop mitigation strategies.

In the hypothetical household, the male head has been HIV-positive for eight years and has had a series of opportunistic infections and illnesses over the past two years that have reduced his contribution to on-farm production by 50 percent and have necessitated his foregoing three months of off-farm work that has traditionally supplemented household income. During the past year his disease has progressed to clinical AIDS and he is no longer able to contribute to productive work either on or off-farm. His wife now devotes 50 percent of her time to caring for her husband and she has had to reduce her agriculturally productive labour by a comparable amount. The couple's four children range in age from 4 to 13 years of age. The oldest child has been withdrawn from school to assist with agricultural production and with child care.

A survey of this household employing human capital indicators should identify a number of impacts on the household's human capital. First is identification of a household member with a chronic illness that limits his productive activities. Secondly, the reduction in labour available to the household suggests a number of likely additional impacts on other dimensions of human capital, as well as other livelihood assets. The withdrawal of the eldest child from school, the reallocation of the wife's time away from productive labour, the loss of off-farm employment, the decline in on-farm labour that is likely to reduce the area cultivated and shift cropping patterns to less labour-intensive practices should all be measured by these indicators. These impacts on the human capital of the household can be expected to directly affect its financial capital. Income from both farm and non-farm sources would be expected to decline. The household will undoubtedly draw upon whatever savings it has and likely call upon remittances from the extended family to help pay for medical care. Depending upon the length of the husband's illness before he dies, the household may also need to liquidate some of its stores of value by selling jewellery, household goods, or other items. Given the total loss of on-farm labour by the husband and the reduction in productive labour by the wife, any surplus agricultural production is likely to be lost and the household becomes oriented toward subsistence and no longer is tied to the market. The household will likely become food insecure in the near future, if it is not already experiencing difficulties.

Other effects may be measurable as well. However, no study can examine all possible impacts of the epidemic. Indeed, we recommend that community-level research using PRA techniques be used to identify those livelihood assets most likely to be affected in the local setting before undertaking a household survey. Detailed household-level surveys should then focus their efforts on the identified dimensions. Failure to focus household surveys on a delimited number of possible impacts weakens the ability of a study to attribute any given effect to HIV/AIDS. The human capital and financial impacts are two of the more important effects in this example. Alternatively, projects that identify natural resources and environmental effects as critical issues might wish to collect detailed information to ascertain if there have been declines in natural capital. For example, the lack of available labour within the household can lead to declines in soil fertility, neglect of on-farm conservation practices such as terracing, and asset stripping as the household attempts to cope with the illness of the head.