Income diversification and poverty dynamics in the rural municipalities of South Africa

Abstract

Poverty, broadly defined as deprivation and lack of opportunities and choices, remains one of the major global challenges. In South Africa, poverty is most prevalent in rural areas. Literature indicates that poverty is dynamic, differs spatially, and households can transition in and out of poverty over time. Income diversification has been identified as a key and common strategy that households use to improve their resilience. In South Africa, the analysis of poverty dynamics has been conducted at aggregated national level without further disaggregation into rural areas, where poverty is most prevalent. Furthermore, household income diversification coping strategies of rural households have not been analysed to determine whether households are diversifying more or less with time, and what the effect of this strategy is on rural household poverty. Understanding poverty dynamics and household income diversification is key for developing policies aimed at reducing rural poverty. The purpose of this study was to investigate income diversification patterns of rural households and their effect on rural poverty across 22 district municipalities in four provinces of South Africa from 2008 to 2017. There are two schools of thought on why households engage in income diversification. The first is that poor households diversify their income out of necessity, desperation, and survival. These negative factors act as “push” factors towards diversification. The second is that diversification of income is used for income growth and accumulation by households with access to assets and high return opportunities. These positive factors act as “pull” factors towards income diversification. This theory underpinned the investigation in this study in rural South Africa. Studies have shown that there are spatial and temporal variations in how households diversify income, whether driven by push or pull factors, as well as variations in poverty dynamics. The spatial disaggregation by rural districts was informed by this literature. A combination of the Simpson Index of Diversity (SID), Foster-Greer-Thorbecke indices, Cox proportional hazard model and an ordered probit model were applied to panel data to investigate the relationship between income diversification and rural poverty dynamics. The data was obtained from the National Income Dynamics Study (NIDS) and covered a period of nine years. The income diversification analysis revealed spatial and temporal variations in household income strategies. This pointed to the importance of disaggregating analyses of household income diversification strategies. Limpopo, KwaZulu-Natal and North West provinces had higher degrees of diversification than the aggregated index, while the Eastern Cape Province had a lower degree of diversification. Contrary to other studies, this study found that provinces with the highest and lowest income did not show the highest degree of diversification. For the low-income households in the Eastern Cape Province, this pointed to entry barriers into high-return activities, while for the high-income households in North West Province, the finding pointed to households that were in general specializing rather than growing their income through diversification. The temporal analysis indicated that these households diversified more over the nine years of this study, with the SID increasing from 0.16 in 2008 to 0.23 by 2017. The poverty dynamics analysis also revealed varied poverty levels across the district municipalities and was most prevalent in Zululand, OR Tambo and Sisonke districts, and lowest in Bojanala, Ngaka Modiri Molema and Joe Gqabi districts. The districts that had the highest poverty rates, also had the highest poverty gap ratios, while those with the lowest poverty rates also had the lowest poverty gap ratios. Poverty transition analyses revealed that, in 18 out of 22 districts (82%), poverty declined between 2008 and 2017, while in 3 districts (14%) poverty increased and in one district the poverty level remained the same. This transition was not mirrored between waves, with the majority of households remaining in the same poverty status between waves (t) and (t+1). This indicated resilience for those households that were non-poor and remained so in the following wave. For poor households, this pointed to a poverty status that did not improve between waves ceteris paribus. The duration models supported these findings, with results indicating that residing in OR Tambo, Amajuba, Sisonke, Ugu, Uthungulu, Greater Sekhukhune, Mopani and Vhembe districts had a reinforcing effect on poverty. Residing in these districts increased the probability of poverty entry and reduced the probability of poverty exit. On the other hand, residing in Umgungundlovu, Capricorn and Waterberg districts increased the probability of a household entering poverty, but had no effect on the probability of exiting poverty. Ngaka Modiri Molema and Zululand districts had the opposite effect. Residing in these districts reduced the probability of exiting poverty but had no effect on poverty entry. Attaining education beyond matric level and job creation in these districts were important for reducing poverty entry and increasing poverty exit. Furthermore, the results indicated that the income diversification strategy was effective at reducing the probability of poverty entry when households had at least three income sources, while increasing the probability of poverty exit when households had at least two income sources. This was important because these districts are predominantly rural, and livelihoods revolved around agricultural activities. The finding pointed to the importance of stimulating the non-farm economy in these districts, as literature indicates the non-farm sector to be a source of high-return activities. The results indicated that income diversification was significant at reducing probability of poverty entry and increasing the probability of poverty exit, thus this strategy should be supported particularly in these districts. A combination of agricultural activities that were already dominant in these districts and non-farm income generating activities could contribute towards this. The study recommends that rural households be supported in their efforts to diversify income as this strategy can improve their resilience by increasing their ability to withstand shocks. In this regard, the recommendation of the study is that provincial governments in KwaZulu-Natal, Eastern Cape and Limpopo target Zululand, OR Tambo, Sisonke, Amajuba, Uthungulu, Greater Sekhukhune, Mopani, Vhembe, Umgungundlovu, Capricorn, Waterberg and Ngaka Modiri Molema districts in their poverty alleviation efforts, specifically by stimulating the non-farm sector and promoting education beyond matric level. Targeting and prioritizing these districts will be important because government resources are limited, and to achieve poverty alleviation will require efficient allocation of these limited resources. This, however, is not to say that other districts be ignored as poverty was also relatively high in those other districts. The recommendation is that more effort be channelled into the identified districts in this study, particularly efforts to stimulate the non-farm economy. The spatial disaggregation and temporal analyses provided insights into the pattern of income diversification and poverty dynamics that might not be observed at aggregated levels. The study contributes to rural household income diversification literature in South Africa by revealing the pattern of this strategy over time and across localities. The study also adds to the poverty dynamics literature, particularly within rural districts, revealing household poverty statuses and transitions over time within these districts. Knowledge of the effect of income diversification on households’ conditional probability of poverty entry and poverty exit is another contribution that the study makes to existing literature. A key recommendation for future research is to further explore the non-farm activities that households in these districts engage in, to better understand rural income diversification. This is because the non-farm sector was found to be relatively important among these households. This could not be undertaken in this research because of data limitations. In addition, future research can extend similar analyses to other rural areas that were not covered in this research.