Malaki Ronald
Mandela

Health shocks are among the most common and least insured risks facing house- holds in developing countries. I examine whether outside household income changes how a caregiver’s labour supply responds to a health shock, using a reporting-bias- purged health stock for adults and children in South Africa’s National Income Dynamics Study and instrumenting household income with the Old Age Pension age-eligibility of a co-resident member other than the caregiver, after showing that the Child Support Grant supplies almost no usable variation because its age ceiling rises at nearly the pace children age between survey waves. A household member’s pension receipt lowers caregiver employment by 7.0 percentage points and hours worked by 8.7 log points. With caregiver fixed effects, the employment effect falls to 4.4 percentage points but remains significant; the hours effect disappears, and I treat it as associational rather than causal. Five of six pension-by-shock interac- tions are null; the exception, a positive interaction between pension receipt and a parental health shock on hours worked, survives a multiple-testing correction and strengthens when shocks are redefined as unanticipated health innovations. The pension relationship does not vary with the household’s chronic health stock, sug- gesting that outside income responds to acute health events rather than to chronic poor health.

Voluntary nutrition labels aim to promote healthier eating by reducing the cost of acquiring nutritional information, yet average effects on dietary quality are small and vary across population groups. We show that this variation depends on the interaction between consumers' internal agency and the external motivations that parenthood introduces. Using five waves of the Household, Income and Labour Dynamics in Australia (HILDA) survey, we combine propensity score matching with a triple-difference design to estimate the post-2014 dietary response of adults with high baseline exposure to discretionary packaged foods, stratified by Locus of Control (LoC) tertile and parenthood status. Among non-parents, high-exposure individuals with middle and internal LoC show dietary quality gains of around 0.17 HEI points after 2014, while externally-oriented individuals show no detectable change. Among parents, we find a large additional dietary improvement for those in the external LoC tertile, a moderate response for those in the middle LoC tertile, and no significant additional gain for those in the highest LoC tertile. The ordering is consistent with the hypothesis that external motivation from parenthood substitutes for internal agency in shaping dietary responsiveness to nutritional information. Formal Wald tests do not reject the null of a common effect across LoC tertiles, so we describe the LoC ordering as a qualitative pattern rather than a formally identified difference. These results suggest that voluntary informational policies reach specific demographic and psychological subgroups, and that their effectiveness depends jointly on the perceived return to health information and life-stage-specific external obligations.

We study the effect of two consecutive Kenyan Pay-As-You-Earn (PAYE) reforms on wage-income inequality using a Fuzzy DiD design with a recentered-influence-function Gini outcome on the Kenya Continuous Household Survey (2019Q1–2022Q4). The 2021 reversal widened the contribution of above-threshold wage observations to the pre-tax Gini by 8.8 percentage points at the central statutory threshold and 20.2 percentage points at the upper threshold, equivalent to 29% and 77% of the local pre-reform Gini at each threshold. The corresponding post-tax effects are statistically zero. The pre-tax widening operates through the formal–informal employment margin, concentrated in Professional occupations and the Public sector; the within-job intensive margin is not detectable. At the top of the Kenyan wage distribution, progressive taxation absorbs the behavioural response rather than further compressing the post-tax distribution. The 2020 cut yields no causally identified effect and is treated as descriptive of the COVID-confounded period.

This paper estimates the relationship between health and labour supply in South Africa using five waves of the National Income Dynamics Study (2008–2017) and a latent health index con- structed from objective, functional, and behavioural indicators. Employment is estimated with a correlated random-effects dynamic probit and hours with system GMM. Both specifications allow the health coefficient to differ by Child Support Grant (CSG) eligibility, gender, and race. A one-standard-deviation increase in health is associated with a 4.5-percentage-point higher probability of employment and with higher hours worked. CSG eligibility is associated with 1.5 percentage points lower employment on average. The association is significantly negative for African adults, indistinguishable from zero for Coloured adults, and significantly positive for White adults. Within the African group the negative estimate is driven almost entirely by women (−12.1 percentage points). Placebo and bandwidth checks around the age cut-off return positive coefficients, so the eligibility association is best viewed as descriptive rather than causal. Health has a clear and broadly shared labour-supply return, while the negative association between CSG eligibility and employment is highly concentrated among African women.