Email us +27 21 650 5696


A key output for each of the REDI3x3 funded projects is the production of a publishable working paper. The papers below are those published by the project, since its inception in the later half of 2012. 

A total of 78 papers were commissioned, most of which are in the process of being finalised. These will be added to the list below, as they are completed.

Martin Wittenberg December 14, 2017

We estimate the parameters of a Pareto distribution for South African earnings as measured through the October Household Surveys, Labour Force Surveys and Quarterly Labour Force Surveys, as assembled in the Post-Apartheid Labour Market Series (PALMS). We develop an outlier detection algorithm consistent with this distribution and then adjust the Gini coefficient for inequality in the top tail, using the robust estimation technique of Cowell and Flachaire. That procedure suggests that wage inequality is a bit higher than conventionally estimated. We also show that the top tail of the South African earnings distribution is ‘thick-tailed’ and explore what that means. Our analyses show big shifts in the distribution in some of the surveys in ways that suggest measurement changes rather than changes in the underlying distribution.

Laura Rossouw, Rulof Burger and Ronelle Burger November 27, 2017

A key factor linked to South Africa’s relatively high maternal mortality ratio is late access to antenatal care. Early access is especially important in the SA context with its high prevalence of HIV amongst pregnant women. This study examined the impact of a package intervention (consisting of an incentive called the Thula Baba Box and a community health worker programme) on the utilisation of maternal health services at an earlier gestational age. To evaluate the impact of the package intervention, a pilot randomised controlled trial involving 100 women of age 18 and older was conducted in an urban area. Women in the treatment group were eligible to receive both the Thula Baba Box and at least two community health worker visits, while women in the control group were subject to standard clinical practice. Both groups were interviewed at recruitment and once again after giving birth. The measured outcomes are the timing of antenatal care visit, whether they attended more than four times and whether they gave birth at a health facility. The paper found that women in the treatment group were likely to seek care on average 1.35 months earlier than women in the control group. They were also significantly more likely to go to the antenatal clinic at least four times. The intervention had no detectable impact on the probability of giving birth at a facility. Women exposed to the Thula Baba Box and community health worker support were significantly more likely to utilise maternal healthcare services than their counterparts. The intervention can be a useful tool to improve maternal health outcomes. 

Rulof Burger and Friedrich Kreuser November 20, 2017

This paper estimates the determinants and effects of latent female household bargaining power for two-adult South African households using cross-sectional data. We apply a new technique which allows us to separately identify relative gender preferences for different commodities and the effect of distribution factors on latent female bargaining power from household survey data. We find that female household members have a stronger preference for expenditure on communication, clothing, personal care and medical expenses, while male members have a stronger preference for alcohol and tobacco, food and entertainment. Additional refutability tests confirm that our two preferred distribution factors – the local sex ratio and the male’s maternal education share – affect consumption decisions via participation in household decisions and not through preferences. Estimates of gendered commodity preferences are used to investigate the effect of other candidate distribution factors on female bargaining power. The results indicate that female bargaining power tends to be higher amongst women who earn relatively more, who come from richer parental households, and who have been married for longer. Finally, our estimates of female bargaining power are used to determine its effect on labour market outcomes. Greater female bargaining power tends to increase the probability that the male will be employed, and that he will work more hours. Although bargaining power does not seem to affect the probability of female employment, employed women with more bargaining power tend to work shorter hours. 

Kezia Lilenstein and Jeremy Seekings November 13, 2017

The high level of youth unemployment in South Africa has prompted investigation surrounding whether youth are holding reservation wages in excess of what they can expect to earn in the labour market, thereby ‘pricing’ themselves out of the market for employment. This paper examines arguments for this theory using data from Cape Town, South Africa. The focus is getting an appropriate measure of the actual reservation wage that features in labour-market decision-making by youth. This paper finds that self-reported reservation wages are unrealistically high for a substantial proportion of the sample when compared with what youth are predicted to earn in the labour market. However, this does not translate into diminished likelihood of employment for these youth. Further analysis indeed reveals that self-reported reservation wages are likely to be too high when compared with the wage at which youth are willing to undertake employment in reality. This is, at least in part, due to the way respondents interpret the question. In addition, self-reported reservation wages may fail to capture the nuances involved in reservation wage formation, in particular job taste considerations made when deciding whether to accept work. This supports the case for constructing a more accurate measure, i.e. of revealed reservation wages, in determining whether reservation wages indeed are unrealistic. When revealed reservation wages of youth are compared with predicted wages, a very small percentage of the youth in the sample are found to be holding unrealistic reservation wages. Therefore, evidence from the CAPS data, utilizing supplementary reservation-wage-inducing questions as well as additional interviews, does not support the claim that unrealistic reservation wages are contributing to the unemployment problem in Cape Town, South Africa. Furthermore, this analysis provides evidence against the standard use of self-reported reservation wage data in the examination of the relationship between reservation wages and employment outcomes. 

Andrew Kerr and Martin Wittenberg October 25, 2017

This paper examines the wage structure and size of the public sector over the post-apartheid period. It does so both descriptively, examining the mean and median earnings in the public sector in the post-Apartheid period and the estimates of total public sector employment and wage bill using household survey data from the post-Apartheid Labour Market Series (PALMS). Trends in earnings and employment from the household survey data are also compared with the macro aggregates published by the South African Reserve Bank (SARB). The earnings premium for public sector workers is estimated using Ordinary Least Squares (OLS) Mincerian wage regressions and quantile regression techniques. This paper also investigates the effects of earnings imputation in the Quarterly Labour Force surveys on the estimates of the public sector earnings premium. 

Martin Wittenberg October 02, 2017

Comparing earnings in the tax assessment data to those in the QLFS, it appears that earnings of employees in the QLFS are underreported by perhaps 40%, with bigger gaps near the top of the distribution. Benefits and annual bonuses contribute substantially to the gap. In the case of self-employment incomes it is also the case that high earnings are missing or underreported in the QLFS, but the tax data seems to miss many mid- and low-income selfemployed earners. These differences make sense when one considers the incentives for reporting accurately to SARS versus to Statistics South Africa. These errors mean that earnings inequality as measured by the Gini coefficient is probably underestimated in the surveys by three percentage points.

Vimal Ranchhod September 29, 2017

How do poor households respond to the cessation of cash transfers in developing countries? South Africa’s generous social pension system results in most of the poor elderly being the primary ‘breadwinner’ in the household. I extract a longitudinal dataset using the rotating panel component of the nationally representative Quarterly Labour Force Surveys, and use fixed effects regression models to estimate the magnitude of changes in household composition and employment that coincide with the departure of a pensioner from the household. I find statistically significant changes in both of these outcome measures. Compositional changes include a decrease in the number of school going aged children, the number of teenagers, and the number of young adults; while the number of older adults increases. I also find significant increases in the number of employed prime aged adults and older adults. The combination of compositional changes and employment changes results in an increase in the mean proportion employed in all of the working age adult groups that we investigate. Overall, households respond by decreasing the number of dependents, increasing the number of potential caregivers, and increasing the proportion of adults engaged in income generating activities. 

Cecil Mlatsheni and Vimal Ranchhod July 19, 2017

The South African youth unemployment rate is extremely high, at well over 50% of youth using the narrow definition of unemployment. This high unemployment rate has been a chronic problem and has been at these or similar levels since the democratic transition in 1994. Unemployment rates were probably also very high for a substantial period before 1994, but the lack of suitable data from that period makes precise measurement of the unemployment rates for that era quite impossible. This paper makes use of the first three waves of the nationally representative longitudinal dataset obtained from the National Income Dynamics Study, to explore the factors that assist school leavers in finding employment. By means of descriptive statistics and regression analyses, it estimates the differences in job finding rates by gender, geographical location, educational attainment, household structure and migration status. The results are not particularly surprising. Male youth, better educated youth, and youth who migrate from rural areas to urban areas tend to have better job prospects on leaving school. 

Vimal Ranchhod and Ihsaan Bassier June 05, 2017

What were the effects of a 52% increase in the minimum wage in the agricultural sector in South Africa in 2013? This paper estimates the short run effects of this policy change on the employment and income of farmworkers, using both repeated cross-sectional data as well as individual level longitudinal data from the Quarterly Labour Force Surveys (QLFS). It finds that the law had a substantial effect on the earnings of farmworkers who remained employed after the law came into effect, but that there was also a small and gradual decrease in agricultural employment. The descriptive evidence from the cross-sections indicates an increase in mean income per month of 17.9% about a year after the law came into effect. This coincided with a mean decrease in adult employment by this industry of about 8.2% over the same time period. Establishing causality empirically is challenging, though. The difference in differences estimates indicate substantial increases in wages in this industry after the law, but this increase is not systematically related to an individual’s wage rate prior to the law. There is also only very limited evidence that employment losses were statistically significant after the law. One explanation for the lack of a systematic relationship between pre-existing wages and subsequent job loss is that the wage gains following the law are observed to be more likely amongst workers who were earning relatively higher wages to begin with. Thus, endogenous compliance or partial compliance may make conventional estimators using a wage gap variable statistically invalid, and may also mitigate against unemployment effects. Overall, the most coherent interpretation of our results is that the law did cause significant increases in income for farmworkers, but did not cause substantial employment losses – although our regression models and data limitations make us cautious about these claims. 

Rulof Burger, Patrizio Piraino and Asmus Zoch May 03, 2017

The paper shows that responses to a traditional, open-ended reservation-wage question are susceptible to a high degree of overestimation and response noise when individuals are rarely confronted with actual wage offers. It argues that, for individuals with weak labor market attachment, a sequence of increasing hypothetical wage offers can more reliably elicit individual preferences: it contains more information about future decisions; it is less sensitive to irrelevant priming effects and more responsive to the economic circumstances of respondents. This has implications for a variety of empirical models in labour economics.