Earnings and employment microdata in South Africa
read more
Citations
The Post-Apartheid Labour Market Series: Social and Behavioural Sciences
Measuring labour earnings inequality in post-apartheid South Africa
The labor market in South Africa, 2000–2017
Separating employment effects into job destruction and job creation: Evidence from a large minimum wage increase in the agricultural sector using administrative tax data
Earnings in the South African Revenue Service IRP5 data
References
Job Flows, Worker Flows and Churning in South Africa
Reweighting South African National Household Survey Data to Create a Consistent Series Over Time: A Cross‐Entropy Estimation Approach
Related Papers (5)
Comparing sample survey measures of English earnings of graduates with administrative data during the Great Recession
Linking ASHE and LFS: Can the main earnings sources be reconciled?
Frequently Asked Questions (13)
Q2. What future works have the authors mentioned in the paper "Wider working paper 2019/47: earnings and employment microdata in south africa" ?
Unfortunately, it does not appear likely that this data source will be made more widely available to researchers and policymakers, after the initial work by Kerr et al. ( 2014 ) documenting the data quality and estimating the extent of job and worker flows. This suggests that admin data are not a panacea, but should rather be used in conjunction with household surveys and firm data to better understand the South African labour market. The authors have tried to document each, focusing on sources that allow researchers to put together a picture of the evolution of earnings and employment in the post-Apartheid period, but pointing out that none are perfect and some may be better than others.
Q3. How many employed individuals are in the recent QES?
The recent QLFSs contain in the sample around 18,000 employed individuals, which is 0.1 per cent of the total employment estimated from this sample.
Q4. What are the main sources of data to be described and analysed?
The household survey data sources to be described and analysed include the Quarterly Labour Force Survey (QLFS) and the older Labour Force Surveys (LFS) and October Household Surveys (OHS), all conducted by Statistics South Africa (Stats SA), starting in 1994 and ending with the most recent QLFS.
Q5. What components were used to estimate the number of informal workers?
In the initial period the authors used four components to estimate informal employment—wage informal, self informal, unpaid family workers, and domestic work.
Q6. What is the reason for the IRP5 data set?
Given the evidence that there is likely to be measurement error in the earnings data from household surveys, the SARS IRP5 data set is thus a potentially very valuable source of data on earnings, since it is believed that it is subject to much less measurement error.
Q7. What is the reason for the decline in agricultural employment?
A recent World Bank report states that agricultural employment halved between 2005 and 2010 (World Bank 2018: 79), but most of this decline is due to the disappearance of the estimated 300,000 individuals working in subsistence agriculture in 2005.
Q8. What did Kerr and Wittenberg (2015) show?
Kerr and Wittenberg (2015) show that this meant that small households were under-represented in the weighted data and suggest that one outcome of this was an undercount of employment in 1998 and earlier, since those in the small households that were missed had better employment outcomes than those in larger households.
Q9. What is the main issue with the QLFS earnings imputations?
The second important issue with the QLFS earnings imputations is that the statistical uncertainty of the earnings numbers obtained from them will be biased downwards, because analysts are treating the data as if they were from actual responses, whereas in reality there is uncertainty about the true values, particularly for those with imputed earnings who refused to answer at all, which occurred from 2010 to 2012 Q2.
Q10. What is the definition of what types of people are considered to be in the informal sector?
Other employed individuals (employers, own-account workers, and persons helping unpaid in their household business) are considered to be in the informal sector if they ‘are not registered for either income tax or value-added tax’.
Q11. How many pension certificates were not identified?
Kerr (2018) did attempt to identify what were probable pension funds, but noted that there were around 500,000 pension income certificates that could not be identified.
Q12. What is the reason for the unreliable period of employment in the SARS IRP5?
That the IRP5 data are a yearly rather than a point-in-time measure and that period of employment seems mis-measured also affect comparisons with household surveys on employment.
Q13. What is the reason why he argues that the unemployment rate increase is understated?
Arrow (2018) incorrectly argues—based on declining realized sample sizes, declining average household sizes, and increases in one-person households—that in fact the unemployment rate increase he documents is understated because of too many one-person households in the samples realized from the 2015 master sample.