Survival strategies of HIV/AIDS – affected households in Soweto

Abstract

This article reports the survival strategies used by households in Soweto with at least one HIV/AIDS-affected member. Data were collected on four occasions between September 2002 and August 2003 from each household, based on diary records and supplemented by interviews with the financial head of the household. In terms of frequency of use and the contribution made to the finances of affected households, the most important survival strategies were: the sick member continuing at work for as long as possible, borrowing from friends and relatives, receiving transfers from relatives, and taking up social grants. These results are broadly similar to those reported by Booysen and associates in their Free State studies.

1. Introduction

The purpose of this article is to identify the specific strategies adopted by a sample of households in Soweto, South Africa, which were directly affected by HIV/AIDS. Soweto is an urban township of more than a million people located adjacent to Johannesburg. We focus on financial strategies, while recognising that a number of other strategies are also likely to be used. This is the first study carried out on income-earning households in South Africa and complements the longitudinal studies of urban and rural households in the Free State of South Africa carried out by Booysen and associates from 2000 onwards.

2. Conceptual framework

We are aware that the concept of the household in Africa is a matter of considerable debate (for a comprehensive review, see Russell, 2003). In the West, the ‘normal’ household is the conjugal nuclear family, which derives from a bilateral descent pathway, whereas African households are based on patriarchal command in patrilineally extended households. This means a much wider range of possibilities for residence in a physical household, as well as for who is expected to contribute to the pool of resources which sustains its residents and who is entitled to use these resources. The pooling of income from multiple sources and its subsequent disbursement is a matter of great practical significance, but was not examined in this study. For practical reasons, we adopted the definition of a household used by Statistics South Africa in their October Household Surveys, i.e. ‘a person or a group of persons who live together at least four nights a week at the same address, eat together and share resources’. We included members who were temporarily absent (e.g. due to hospitalisation) and heads of households who were migrant workers. The latter were included because of their importance in household earning and spending decisions.

Previous studies of household coping or survival strategies in the face of HIV/AIDS have almost all been carried out in rural households and communities. Sauerborn et al. (1996), for example, studied the strategies used by individuals and households in Burkina Faso to cope with the economic costs of illness; the specific illnesses were not identified but were classified as mild/moderate or severe/fatal. The authors developed a taxonomy of 11 types of coping behaviour, including attempts to avoid costs by ignoring a disease and then progressively attempting to minimise the impact on the household once the illness was perceived (e.g. by mobilising savings, selling assets, employing wage labour and offering wage labour to other households). The study found intra-household labour substitution and sale of livestock to be the most common strategies and that coping was very largely a household undertaking; inter-household transfers were very limited. Mutangadura et al. (1999: 17) reviewed survival strategies for rural households and communities, as did Rugalema (2000: 539). These studies concentrated on describing strategies and reporting their frequency of use and sequencing, rather than assessing their effectiveness; however, some (e.g. Sauerborn et al., 1996) did examine whether the strategies were broadly successful (e.g. did they enable the household to survive?).

Of particular relevance is the comprehensive study of the impact of HIV/AIDS on households in rural and urban areas in the Free State of South Africa by Booysen et al. (2004). The research began with 406 rural and urban households in mid-2001 and concluded with 351 households at the end of 2002, although there was some further data collection up to the end of 2004. Given the study's hypothesis, which was confirmed by its data, that the presence of a member with HIV/AIDS is likely to push a household towards or into poverty, one of the study's aims (2004: 18) was ‘to describe and evaluate the impact of different informal coping strategies and support systems adopted by individuals, households and communities, as well as that of formal HIV/AIDS-related interventions of national and provincial government departments and local authorities, in terms of their impact over time on the quality of life of affected households’. The most important strategies investigated in their research were the migration of individuals into and out of households, financial strategies (borrowing, use of savings and sale of assets) and access to social grants. Comparison of our findings with those of Booysen et al. are made at appropriate places.

The insights of these studies were combined to produce a taxonomy of survival strategies under four broad headings (see Table 1):

• strategies aimed at alleviating the loss of income

• strategies aimed at surviving the financial cost

• strategies aimed at alleviating the loss of labour

• strategies aimed at using safety networks

Table 1: Household survival strategies

Strategies aimed at alleviating the loss of income	Strategies aimed at surviving the financial cost	Strategies aimed at alleviating the loss of labour	Strategies aimed at using safety networks
Encouraging employed people to join the household	Utilising savings	Reallocating household labour	Inter-household care and psycho-social support
	Disposing of assets
Diversifying income	Borrowing	Encouraging non-employed members to look for work	Inter-household financial support
Remaining at work until the person can no longer work	Taking children out of school	Withdrawing children from school	Fostering of children
Dissolving the household	Reducing consumption	Working extra hours	Transfers in from relatives and non-relatives
	Begging
Doing nothing	Adjusting household size and composition		Coping with discrimination
	Applying for state assistance

3. Methods

In our study, the sample households were identified and recruited in the second half of 2002, using a purposive sampling procedure. Potential households were identified from women who attended public antenatal clinics in Soweto or the HIV clinic at the Chris Hani Baragwanath Hospital. Thus a limitation of the study is that it does not include any HIV-affected people who did not visit a clinic. Public health practitioners assisted in selecting the sample of households, based on the willingness of patients to be tested for HIV and to participate in the study. Each household had to have at least one member engaged in an income-generating activity. Affected households included a member with a CD4 cell count of 200 or less. The study began with 125 households, of which 113 (61 in the affected cohort and 52 in the non-affected cohort) were followed between September 2002 and August 2003. Data collection was based on a diary prepared by the first author and completed by the person in each household who was responsible for its finances. Diary entries were checked, corrected and amplified during four visits over the 12 months by fieldworkers who were trained and supervised by the first author. Full details of the diary and data collection procedures, along with complete results, are presented in Naidu (2004).

Methodological purists may point out that the households in this study were not chosen by random methods and that the results may therefore be suspect. However, it is almost impossible to use randomly selected households in this type of study. To be pure, we would need to take several hundred households, test the HIV-status of each of their members and exclude any household with a HIV-positive member. We would then need to follow the remainder over, say, ten years and engage in regular testing of HIV status. If a household member became HIV positive, their household would become part of the experimental group. This would enable ‘before’ versus ‘after’, and ‘with’ versus ‘without’, comparisons. The ethical, time and financial constraints obviously make such a method unusable. We need to accept less pure methodologies and hope that these will provide a reasonably accurate picture of the consequences of HIV/AIDS for households. The study by Booysen et al. (2004: 28), which is also based on purposive sampling, makes similar admissions about methodology. However, if a number of methodologically impure studies come up with similar findings, then we can have some confidence that we are on to something. In recognition of the study's methodical impurity, we use the terms ‘affected’ and ‘non-affected’ households rather than ‘experimental’ and ‘control’ groups.

It might be questioned whether the apparent responses to HIV/AIDS which we identify are in fact a consequence of differences between the households, for example in size, age composition or economic circumstances. It is difficult to prove that the two groups of households were similar before the advent of HIV/AIDS because the affected households had already experienced some of its negative consequences before the study began and it was not possible to secure retrospective data in sufficient detail. In some important respects, however, including their size and age composition and the value of their marketable assets (very largely housing), the households were very similar. Generally speaking, we believe that the differences we identify between affected and non-affected households are an accurate reflection of the economic impact HIV/AIDS has on households and on the survival strategies which the affected households use. This belief is strengthened by the longitudinal aspect of the study, which allowed observation and documentation of changes over a 12-month period.

4. The economic impact of HIV/AIDS on the sample households

The impact of HIV/AIDS on affected households' income and expenditure patterns is the subject of an article currently in preparation and is dealt with briefly here. The study distinguished between regular and irregular income. Regular income included earned income (from employment or self-employment) and income from non-market sources (from social grants or monthly transfers from people outside the household). Irregular income comprised earnings from property and investments (including rental and sale of assets), from non-market sources (such as child maintenance, irregular transfers from people outside the household and profit from the sale of home production) and income as a result of death. The last item included lump sum payments or income from burial societies, relatives and non-relatives. The incomes of affected and non-affected households over the year are reported in Table 2.

Table 2: Annual household income by source (rand, mean adult equivalent)

	Affected households (61)	Non-affected households (52)
Regular income
Earned	7 540	10 940
Non-market sources	4 527*	1 604
Irregular income
Property, investment	227	53
Non-market sources	509	407
As a result of death	1 703	145
Total	14 556	13 149
*Of which R4381 (96.7 per cent) comprised social grants.

Using mean annual adult equivalent figures, affected households received 31 per cent less earned income than non-affected households. This was more than compensated for, however, by affected households receiving 65 per cent more regular income from non-market sources, in particular social grants, and significant irregular income as a result of death. Total income received by the two groups was not significantly different. Social grants were very largely disability grants, old age pensions and child support grants. The proportions of affected households receiving these grants (24, 18 and 22 per cent respectively) were much greater than the proportions for non-affected households (4, 8 and 12 per cent, respectively). For affected households, these contributed 30 per cent of annual household income compared with 11 per cent for non-affected households.

HIV/AIDS can reduce regular income flows into a household in at least four main ways:

• temporarily, as a result of absence from work by a HIV-infected worker

• temporarily, and possibly permanently, through loss of earnings by caregivers

• permanently, when income that has been temporarily increased from a disability pension granted to a person with HIV/AIDS ceases at that person's death

• permanently, through loss of employment as a result of sickness and/or death.

Table 3 reports the frequency of these effects, which explains the finding reported in Table 2 that affected households earned 31 per cent less than non-affected households. Absence from work did not always result in lost earnings, as about a third of absences were supported by sick leave or worker's compensation. However, absences as a result of HIV/AIDS led to subsequent loss of employment. By far the largest cost associated with caregiving was the opportunity cost of foregone housework and gardening.

Table 3: Frequency of income reducing effects (number of households)

	Affected households (61)	Non-affected households (52)
Sick member absent from work	17	4
Caregivers absent from work	9	1
Loss of employment	12	1
Loss of social grant	9	1

As for expenditure, Table 4 shows that over the 12 months more than three quarters of the sample households' expenditure was made on regular monthly categories. The ‘basic essentials’–housing (rent, electricity, water), food, non-food groceries, clothing and public transport–made up about half of total expenditure There were no significant differences between affected and non-affected households with regard to regular expenditure, expenditure on basic essentials or total expenditure.

Table 4: Annual expenditure patterns of sample households (rand, mean adult equivalents)

	Affected households (61)	Non-affected households (52)
Regular expenditure	7 788	8 541
Basic essentials	5 336	5 486
Irregular expenditure	3 245	1 603
Health care	341	32
Funerals	1 883	355
Total expenditure	11 033	10 144

There was a significant difference in irregular expenditure, which is very largely explained by much higher health care and funeral costs incurred by affected households. These made up 18 per cent of this category's total expenditure compared with 4 per cent among non-affected households.

To sum up: affected and non-affected households received similar total incomes, but affected households earned less regular income and were heavily dependent on social grants. Both groups of households spent similar total amounts, and similar amounts on basic necessities. Affected households spend significantly more on health care and funeral costs. In their Free State study, Booysen et al. (2004: 76–7, 80) found that mean adult income per capita in affected households was 62 per cent of those in non-affected households; the corresponding figure for expenditure was 67 per cent. Their study was carried out before the dramatic increase in the take-up rates of child support, disability and foster care grants which began in 2004; the much higher take-up rates among the Sowetan households appears to explain why total income and expenditure levels were similar in affected and non-affected households. Regular earned income for affected households in our sample (see Table 2) was 69 per cent of that for non-affected households.

5. Household survival strategies – results and discussion

5.1 Strategies to minimise the loss of income

Sick household members who were employed almost always kept on working until they were unable to continue. Apart from the fact that wages exceeded the disability pension, higher benefits were received if the sick person died while still employed, as opposed to dying following early retirement.

In 30 per cent of affected households, one or more members tried to increase household income in response to HIV/AIDS, very often by selling items such as cigarettes and sweets, or renting telephones. Given that the unemployment rate in Soweto was 57.3 per cent (Naidu et al., 2004), the chances of unemployed people finding work were low. There was little scope for employed household members to take on a second job, given that they already worked an average of 8,5 hours a day, with another 2,5 hours spent in travelling.

5.2 Financial strategies

The financial situation of households at the start of the study period is shown in Table 5. Affected households began the study period with 69 per cent more debt than non-affected households, 49 per cent less in savings (although savings by all households were small) and similar values of marketable assets. The last consisted very largely of housing. By the end of the study period, the affected cohort had larger savings levels, largely as a result of lump sum receipts following deaths.

Table 5: Financial situation as at September 2002 (rand, mean)

	Affected households (61)	Non-affected households (52)
Indebtedness	13 309	7 871
Savings	177	234
Value of marketable assets	45 536	46 492

Table 6 lists the proportion of households which used any of three common ways of finding additional finances during the study period – borrowing, dissaving (the use of savings to fund spending that is greater than income) and selling assets. The Free State study (Booysen et al., 2004: 82–99) found the same order of priority: affected households would first borrow, then deplete savings and then, as a last resort, sell assets.

Table 6: Frequency of use of financial strategies (per cent of households)

	Affected households (61)	Non-affected households (52)
Borrowing	64,7	50,9
Dissaving	11,8	33,9
Selling assets	10,3	3,4

Borrowing was by far the most common financial strategy and was more commonly used by affected households. The main purpose of borrowing was for food, transport and funerals, and the main sources were relatives or friends (60 per cent of loans for all households) and micro-lenders (15,8 per cent). Booysen et al. (2004: 92) found similar frequencies in the Free State. Affected households borrowed a good deal less than non-affected households (a mean of R1 196 compared with R2 862), which is perhaps not surprising given their high initial levels of debt. Affected households also allowed the amounts owed for past water and electricity consumption to accumulate.

Similarly, affected households withdrew less frequently (on 17 occasions compared with 48) from their savings and withdrew smaller amounts (a mean of R330 over a period, compared with R442). Most savings were in the form of burial society funds and were not therefore readily accessible. Withdrawals were used for purposes very similar to borrowing. Selling assets seems to have been a last resort. Nine affected and two non-affected households sold assets but the proceeds were small relative to borrowing and dissaving. The total proceeds from selling assets averaged R19 for affected and R10 for non-affected households, respectively.

The importance of social grants in the financial survival of affected households was noted in Section 2, with 30 per cent of the affected households' income coming from social grants, compared with 11 per cent from non-affected households. Old age pensions are payable to males aged 65 and over and females aged 60 and over and are subject to a means test. The maximum qualifying annual income, after certain deductions, was R19 080 for a single person and R31 320 for a married couple. The take-up rate among the eligible population in the sample was 100 per cent. In 2002, the Child Support Grant was available on behalf of children under the age of seven (this rose to age nine in 2003). The take-up rate by eligible households was higher among affected households than among non-affected households (72 per cent of households compared with 61 per cent). The Disability Grant was available only to those who produced a medical report stating that the individual was unable to provide a livelihood. The take-up rate of 94 per cent is not surprising because of the sampling methodology used in this study: in all of those identified as HIV positive, the infected individuals chosen had an initial CD4 cell count of 200 or below, which is the benchmark used by medical practitioners to decide on disability status. Those infected people not receiving the disability grant were either in the process of applying for it or waiting for their CD4 cell counts to drop further in order to do so. Booysen et al. (2004) also reported the crucial role of social grants in the survival of affected households, but suggested (2004: 78–9) that social grants may to some extent be crowding out private transfers, including remittances.

5.3 Using social networks

Affected households drew on a range of social networks as part of their survival strategies. We have seen that relatives and friends were the major source of lending to affected households. Funerals were heavily subsidised by financial and other contributions from relatives and friends. In the past few years, monetary contributions from neighbours have become much less common, which respondents explained in terms of the greater frequency and high cost of funerals.

Affected households were more likely to receive monetary transfers from relatives elsewhere. Some 74 per cent of affected households received such transfers, compared with 42 per cent of non-affected households. However, there were transfers into and out of all households and the net inflow into affected households amounted, on average, to only R141 during the year under study. Non-affected households had a net outflow of R727. The amounts received as a result of death were much greater. Table 2 showed that affected households received 11,7 per cent of their annual income as a result of death, which included payments from burial societies. Overall, transfers from other households provided only limited financial resources to affected households, although the timeliness of its arrival gave it an increased value.

5.4 Movement of household members

Households in both cohorts sent members, usually children, to be taken care of by other households, and this was more common in the affected cohort. However, very few children – a total of five – were sent away as a result of illness or death. Eight sick individuals were sent to live elsewhere. Booysen et al. (2004) reported much higher migration rates: 235 people (12,3 per cent) of their study population migrated from households and 184 people migrated into households during the 18 months of their study. The most important motive given for out-migration was employment (a quarter of all households) while the most important reason given for in-migration was to take care of an adopted or fostered child (2004: 56, 63). While these main motives were similar for affected and non-affected households, 40 per cent of affected households sheltered an orphan during the study period, compared with a quarter of non-affected households.

Finally, some potential strategies were not reported by any of the affected households in the sample. These included taking second jobs, begging and taking children out of school in order to save school fees or to use the child's labour in the household. There was some evidence of taking children out of school in the Free State study (Booysen et al., 2004: 117–8), where 9,9 per cent of children from affected households did not attend school compared with 4,8 per cent of children from non-affected households.

6. Conclusions

The most important survival strategies identified in this study were the continuation at work of the sick member for as long as possible, borrowing from friends and relatives, receiving transfers from relatives and taking up social grants. In terms of the frequency of their use and the contributions they made to household finances, these were by far the most important survival strategies.

In determining the success of the coping behaviour in rural contexts, Sauerborn et al. (1999: 292) suggest three questions: ‘First, did coping [strategies] avert calamity – i.e. the social and economic breakdown of the household? Second, did it reduce or avert any negative effect of illness on household production? Third, did coping strategies preserve household assets?’ If we modify these to fit an urban context, the experience of the Sowetan households suggests a positive answer to each question. In particular, only three households dissolved during the study period. In each case, the reason was the death of an adult such that only children remained in the household and these were relocated to households outside Soweto. The Free State study did not examine the dissolution of households, although 55 of the original 406 households could not be studied throughout the period, either because they moved outside the study area or their whereabouts could not be established (Booysen et al., 2004: 23, 49–50); some of the latter may have been dissolutions.

However, Rugalema (2000) makes a strong critique of the concept of coping strategies in respect of HIV/AIDS, again from a rural household perspective. He argues that coping implies that after some major setback a household is able, by using appropriate strategies, to return to its former standard of living or better, and attributes the popularity of the concept of coping to the free market worldview. That is, when confronted with a disaster, individuals and households are able to engage with the market (e.g. by selling assets) in order to deal with the situation. Rugalema questions whether households indeed have rational plans and strategies during times of disaster; rather, he suggests, they ‘they react to the immediacy of need, disposing of assets when no other alternatives present themselves’ (2000: 541). We found substantial evidence of a lack of planning among Sowetan households (e.g. few adults had made wills and few parents had made arrangements for the care of their children in the event of their death). The long-term consequences of a disaster such as HIV/AIDS mean that households are likely to emerge with a substantially lower standard of life than before (Naidu & Harris, 2005), unlike, for example, recovery following the effects of drought. To Rugalema, therefore, HIV/AIDS affected households are best described as struggling to survive rather than coping.

We can note several policy implications from this study of survival strategies. First, following the argument of the previous paragraph, belief in the efficacy of household coping strategies can give the government an excuse to do little to alleviate the effects of HIV/AIDS on communities. For example, the South African government's emphasis on home-based care in fact shifts the burden of care to struggling households and communities. Secondly, social grants have been crucial, as has the financial and broader support of relatives, in compensating for the lower incomes and higher expenditures which HIV/AIDS imposes on households. Further analysis of the data used has found that less than 10 per cent of affected households fell below the poverty line, as measured by an expenditure of R250 per adult equivalent member per month (Naidu & Booysen, 2005); there was no significant difference between affected and non-affected households in this respect. In the context of massive numbers of premature deaths as a result of AIDS, however, financial support from neighbours to affected households has diminished. We have reported elsewhere (Naidu & Harris, 2005) that the average cost of funerals to affected households, when pre- and post-funeral ceremonies are included, is around R32 000. It is possible that support from extended family members may begin to reach some limits in the near future. If so, social grants will become even more important to affected households.

This research was financially supported by USAID, DFID and AusAID, and administered by the Joint Economics, AIDS and Poverty Programme (JEAPP) under a subcontract from Nathan and Associates Inc.