Economic and health implications of tobacco and alcohol use in South Africa: a household expenditure analysis

Chisha, Zunda

Economic and health implications of tobacco and alcohol use in South Africa: a household expenditure analysis

Thesis / Dissertation

2025

Abstract

This study uses nationally representative data to analyse the economic and health implications of tobacco and alcohol use in South Africa. Issues covered include the crowding-out effect of tobacco expenditure on household expenses, and the impact of parental tobacco and alcohol consumption on child health outcomes. The thesis emphasises the interconnectedness of tobacco and alcohol use, and their place in South Africa's development and policy landscape. Chapter 2 uses four income and expenditure surveys to investigate the categories of household expenditure displaced by tobacco expenditure. The analysis begins with a simple comparison of differences in expenditure shares between households that spent money on tobacco and those that did not. Next, the displacement of household expenditures is estimated using instrumental variables within a system-of-equations. Confidence intervals for the point estimates are then estimated using a method proposed by Nevo & Rosen (2012). Results indicate that expenditures on education, transport, housing, and household goods were displaced by tobacco expenditure, however, alcohol expenditure was sometimes crowded-in. Decreased household spending on education could lead to reduced human capital development. Chapter 3 looks at how crowding-out patterns changed as tobacco taxes increased in the period 1995-2005/6. It uses a semi-parametric approach for the difference-in-difference technique developed by Abadie (2005), accounting for violations of the parallel-trends assumption. The results suggest a non-uniform pattern of adjustment of expenditures in tobacco-consuming households. While budget shares increased for some categories, such as food and alcohol, others, such as education, were crowded out. The displacement of expenditure on education is proportionately greater for low-income households than for those with higher incomes. Chapter 4 looks at whether parental tobacco and alcohol impact child health outcomes by crowding-out household food expenditure. Using instrumental variable estimation and mediation analysis, the study uses the National Income Dynamics Study (NIDS) panel datasets to investigate the lagged long-term effects of tobacco and alcohol expenditures on two child anthropometric measures (height-for-age and weight-for-age z-scores). Although a direct link between reduced food expenditure and observed effect is not established, tobacco and alcohol expenditure are shown to have a substantial negative impact on child anthropometric measures. The study has three main implications. First, the dynamic analysis shows how various factors interplay with changes in fiscal policy, affecting household-level budgeting decisions. In doing 5 so it helps explain the effectiveness of cigarette tax policies over time, highlighting their continued relevance. Secondly, the study reveals possible unintended consequences among households that continue purchasing cigarettes after substantial price increases, such as deeper expenditure displacements, which policymakers must consider. Lastly, the evidence generated here suggests that reductions in tobacco use can be associated with a reduction in alcohol use, thus tobacco control could be used a tool for alcohol control.