Abstract
Vulnerable groups of people experience numerous socioeconomic challenges, such as poverty, hunger, unemployment, poor quality of education and HIV/AIDS, due to the unequal allocation of resources in a precarious social, historical and economic setting like South Africa. These issues pose a major challenge in ensuring comparable levels of well-being, especially among older people, smallholder farmers and women. For instance, about 5.51 million older people struggle with their health, financial security, and social isolation. Additionally, approximately 2.4 million smallholder farmers face poverty, food insecurity, climate change, and economic and infrastructural constraints while lacking targeted welfare programmes. Furthermore, an estimated 31 million women experience compounded discrimination worsened by unprecedented challenging times like COVID-19. For policymakers to effectively address specific issues, such as those mentioned above, accurately measuring these vulnerable groups' well-being is imperative.
Although many international studies have investigated the well-being of various vulnerable groups of people, we contribute to the understanding of their vulnerability in the following ways. Firstly, we highlight the need for a micro-level multidimensional composite quality-of-life index for older people. Compared to the current international indices, it improves the granular investigation of older people's concerns. Additionally, it allows us to follow the same cohort while determining over time trends and obtaining the demographics of the most vulnerable within the older population. Secondly, while previous studies concentrated only on a single welfare payment when investigating the needs fundamental to the subjective well-being of smallholder farmers, we analyse two types of payment: an earmarked grant (land grant) and unconditional social welfare payments. Furthermore, we explore how food insecurity, proxied by crop failure, ties in with the welfare payments necessary to enhance the well-being of smallholder farmers. Thirdly, we follow the same cohort of women over time to capture how their mental health changed during unprecedented times of hardship and determine those factors that contributed to their vulnerability—something previous studies could not do given data limitations.
Consequently, this thesis aims to assess the levels of three distinctly different well-being constructs (objective, subjective and mental health, as determinants of well-being) among older people, smallholder farmers and women in South Africa. The first objective is to build a unique
vi
quality-of-life index focusing on older people. The index allows us to identify those dimensions that explain the most variance in the quality of life of older people, determine the trend and compare the quality of life of different older demographic groups. The second objective is to identify whether the relationship between welfare payments (both social and land) and life satisfaction is influenced by crop failure. Specifically, it investigates whether crop failure affects the strength, direction and underlying process by which social welfare impacts the life satisfaction of smallholder farmers. The third objective is to assess women's mental health during the first year of the COVID-19 pandemic, how it changed, and the significance of the pre-existing (prior to the pandemic) depressive symptoms.
The first objective was accomplished by constructing the first multidimensional quality-of-life index for older adults in South Africa at a micro-level. We functionalised the index using the South African National Income Dynamic Study (NIDS) dataset. Our results from Categorical Principal Component Analysis (CATPCA) show that the dimensions explaining the most variance are household services, economic status, safety, and mental and physical health. The well-being of older adults increased over time, likely due to policy interventions, with the most vulnerable group being black African older women in traditional and farming areas.
The second objective utilised the NIDS and the South African Weather Service datasets. A pooled ordered probit model showed that social welfare payments (SWPs) decreased the life satisfaction of smallholder farmers. In contrast, the agriculture-specific government welfare payment (WP) 'land grant' increased life satisfaction. Moreover, when a household experienced crop failure and received an SWP, it reduced the life satisfaction of smallholder farmers. In terms of agriculture-specific government WPs, i.e., land grants, we found that if crop failure occurred and a household received land grants, it increased the life satisfaction of smallholder farmers; however, the relationship was not statistically significant. Additionally, the causal mediation analysis showed that crop failure explained 4% of the negative relationship between SWPs and the life satisfaction of smallholder farmers (other factors explain the rest). However, crop failure did not contribute to explaining the process of how land grants increased smallholder farmers' life satisfaction. Therefore, neither an unconditional SWP nor an earmarked land grant (neither specific to food security) had the necessary effect on food production.
The third and final objective utilised the NIDS and the Coronavirus Rapid Mobile Survey (NIDS-CRAM) dataset to assess women's mental health vulnerability, which was already a
vii
concern before the COVID-19 pandemic. We found that the value of the depressive symptoms scale significantly increased during the first period of the pandemic and then eased over time. Interestingly, the behaviour of the individual scale items ('depressive feelings' and 'anhedonia') differed over time. This result questions the internal reliability of the scale during the pandemic and the importance of analysing the mental health scale items individually. Furthermore, a pooled ordered logit indicated that being depressed before the pandemic increased the probability of 'depressive feelings' and does not matter for 'anhedonia'. Other factors, from the fixed-effects ordered logit models, found to be increasing the probability of mental health disorders were taking care of children for 13-24 hours a day and living with a person who has gone hungry. In contrast, wearing a mask and living in a grant-receiving household decreased the probability.
Given our results, additional specific interventions are needed to improve the well-being of the vulnerable groups. For older adults, possible policy initiatives could focus on improving housing conditions, reducing crime rates, improving access to health care, and creating tailored mental health policies. For smallholder farmers, we recommend a more targeted approach to welfare payments focusing on the risks of the occurrence of agricultural losses, enhanced crop insurance availability, access to pest control, public irrigation systems, and training in sustainable farming practices. Regarding women's well-being, especially for those with pre-existing psychological vulnerabilities, there is a need for mental health policies, in the form of improvements in employment conditions such as job security and paid sick and maternity leave. Introducing these more specific initiatives parallel to the existing policies, can provide for basic human needs and improve the well-being of vulnerable groups in South Africa and the rest of the developing world.