Gender inequality in the South African labour market: Insights from the Employment Equity Act data

abstract

In the years following the end of Apartheid, South Africa introduced the Employment Equity Act as part of legislation designed to address entrenched racial and gender inequalities. Through section 27 of the Act, firms are required to report on the representation and remuneration of their workers by gender and population group. In this study, unprecedented access to the data generated by this reporting was used to assess gender-based occupational segregation and pay gaps in 2015 and 2016. We found the data on employment to be relatively trustworthy and to show that women (and especially black and coloured women) continue to be under-represented in high-skilled and management positions. Substantial gender pay gaps were found, but the reliability of these estimates was made questionable by a high number of apparent errors and inconsistencies in the remuneration data. The data has several advantages over other household and firm surveys, but these issues undermine its potential for the estimation of gender pay gaps among workers matched on occupational skill levels. This analysis shows that more needs to be done to ensure that this data is appropriately processed and distributed, so that it can shed light on the state of women in the South African labour market and be used to effectively address inequalities in pay and representation.

keywords:

• pay
• inequality
• women
• gender
• labour
• equity

Notes

1 The remuneration data is comprised of any payment in money and in kind, including bonuses and share options, and full-time equivalent amounts are provided for temporary and part time workers (Department of Labour, 2014).

2 The ECC has now been superseded by the National Minimum Wage Commission.

3 The ECC has, so far, failed to investigate norms and benchmarks for proportional income differentials, and there has been no action by the Minister of Labour to reduce income inequality under the act (Helm, 2014). Furthermore, only a handful of cases of firms being prosecuted for noncompliance have been recorded, all due to firms not submitting EE reports, despite the Act detailing maximum fines (which were increased in 2014) that firms can be subject to (van Rensburg, 2017). A lack of clarity in the EEA surrounding its legal implementation may have contributed to this (Bezuidenhout et al, 2008).

4 While the employment estimates of the QES are typically lower than those of the QLFS, its estimates of average income are generally far higher. Wittenberg (2014) unpacks the differences between the QLFS and QES in terms of both employment and average income. He finds that a large part (15%) of the gap in average income estimates remains unexplained even after adjusting for various factors surrounding coverage (e.g. omitting self-employed and agricultural workers in the QLFS) and misclassification. He also notes concerns about the lack of representativity of the frame from which firm surveys are drawn.

5 The data comes from EEA4 forms submitted by designated firms to the Department of Labour, and which the Department has made available for this research. Other information collected by the forms covering business type and province were not available in the data.

6 The possible reasons provided are employees’ underestimation or ignorance of the number of employees in their place of work (thus reducing the sample), or because of an underestimation of the South African population in the weights used in PALMS to generate population estimates.

7 If the remuneration per person in 2016 for any group (defined by sector/occupational level/gender/race) was more than double or less than half the remuneration per person for that same group in 2015 this group was flagged. Thereafter we compared the discrepancy of each year’s value from the mean for all groups in the same category (by sector and occupational level) to locate the likely source of the error (2015 or 2016). It must be borne in mind that the pay gaps described could thus be either underestimates or overestimates depending on the characteristics of omitted groups. Of the 173 errors identified (in 1540 total groups), the number of errors was fairly evenly distributed across different occupational levels. More errors were found among women than men, and more among black and Indian groups than white and coloured ones, and there were variations in the accuracy of the data between sectors. There was no clear way to deal with these problematic observations without raising other issues. Omitting values based on some rule of thumb runs the risk of masking some significant changes from year to year and potentially underestimating the severity of gaps, while also introducing the risk of bias (should sample characteristics determine the likelihood of inclusion in the final sample). Besides omitting errors, we also experimented with a strategy of replacing values identified as errors with inflation-adjusted values from another year. This imputation process has the advantage of maximising the sample and should not be influenced as much by the specific characteristics of the omitted groups, but as when taking the data at face value this strategy resulted in a string of unintelligible estimates and patterns.

8 The lack of hours worked information makes it difficult to compare temporary employees to other workers in terms of annual remuneration.

Additional information

Notes on contributors

Gabriel Espi

GABRIEL ESPI has studied at the University of the Witwatersrand and the University of Cape Town and has worked in research and teaching at both institutions, including the Southern Centre for Inequality Studies at the former and Data First at the latter. His interests include the psychology of language, the area in which he completed his Master of Arts degree - focusing on the effect of multilingualism on the mind - and labour and health economics, covering topics such as the position of women in the economy and child welfare. Email: [email protected] (http://orcid.org/0000-0003-1494-9308).

David Francis

DAVID FRANCIS is the Deputy Director at the Southern Centre for Inequality Studies, and a PhD candidate in Economics at Wits University. His research interests focus on labour market economics, the informal economy, and inequality. He was the researcher for the Advisory Panel on the minimum wage in South Africa. He has previously worked as a development consultant, and a policy and budget analyst at South Africa’s National Treasury, where he worked in health and social policy. He has an MA in development studies from the University of KwaZulu-Natal, and a bachelor of social sciences in economics and history from the University of Cape Town. Email: [email protected]

Imraan Valodia

Professor IMRAAN VALODIA, an economist, is Dean of the Faculty of Commerce, Law and Management at Wits University. His research interests include inequality, competition policy, employment, the informal economy, gender and economic policy, and industrial development. Imraan has led the initiative at Wits to establish the Southern Centre for Inequality Studies - a multi-disciplinary, cross-country initiative to promote research and policy change to promote greater equality in the global South. In addition to his duties as Dean, Imraan leads the SCIS. Imraan has led and participated in a number of large national and international studies. He is recognised nationally and internationally for his research expertise in economic development. Imraan is a part-time member of the Competition Tribunal in South Africa. He is also a commissioner of National Minimum Wage Commission, and a member of the Academy of Science of South Africa (Assaf) Standing Committee on Science for the Reduction of Poverty and Inequality. In August 2016 Professor Valodia was appointed by President Cyril Rampaphosa to chair the Advisory Panel on the National Minimum Wage. This led to the introduction of a National Minimum Wage in South Africa. In early 2018, he was appointed to a Panel to advise the Minister of Economic Development on amendments to the Competition Act, which were introduced into law in 2019. He was also appointed by the Minister of Finance to a Panel of Experts to advise on Value-Added Taxes. Email: [email protected]