The Determinants of Earnings Inequalities: Panel Data Evidence from South Africa

36 Pages Posted: 12 May 2012

See all articles by Andrew Kerr

Andrew Kerr

University of Cape Town (UCT)

Francis Teal

University of Oxford - Department of Economics

Abstract

In this paper we analyse the relative importance of individual ability and labour market institutions, including public sector wage setting and trade unions, in determining earnings differences across different types of employment. To do this we use the KwaZulu-Natal Income Dynamics Study data from South Africa, which show extremely large average earnings differentials across different types of employment. Our results suggest that human capital and individual ability explain much of the earnings differentials within the private sector, including the union premium, but cannot explain the large premiums for public sector workers. We show that a public sector premium exists only for those moving into the public sector. The paper addresses the challenges of non-random attrition and measurement error bias that panel data bring. Our results show that emphasising a simple binary dichotomy between the formal and informal sector can be unhelpful in attempting to explore how the labour market functions.

Keywords: formality, trade unions, public sector, earnings, South Africa

JEL Classification: J31, J51, J45, O12

Suggested Citation

Kerr, Andrew and Teal, Francis, The Determinants of Earnings Inequalities: Panel Data Evidence from South Africa. IZA Discussion Paper No. 6534, Available at SSRN: https://ssrn.com/abstract=2056724 or http://dx.doi.org/10.2139/ssrn.2056724

Andrew Kerr (Contact Author)

University of Cape Town (UCT) ( email )

Private Bag X3
Rondebosch, Western Cape 7701
South Africa

Francis Teal

University of Oxford - Department of Economics ( email )

Manor Road Building
Manor Road
Oxford, OX1 3BJ
United Kingdom

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
162
Abstract Views
797
Rank
333,835
PlumX Metrics