Skip to content

Unemployment in South Africa: Descriptors & Determinants

Literature › Works

Haroon Bhorat (2007). Unemployment in South Africa: Descriptors & Determinants. Presentation to the Commission on Growth and Development, World Bank, Washington DC, 15 October. School of Economics & Development Policy Research Unit, University of Cape Town.
Type conference paper
Scope Post-1994 South African economy
Method mixed
Period studied 1990–2005
Regions South Africa
Themes inequality poverty unemployment, labour markets wages, education human capital, unions strikes, transition post apartheid, demography urbanisation
Citation key bhorat2007b

Argument

The slides describe post-apartheid unemployment using household survey data for 1995-2005 and a multivariate probit analysis. They argue that race, gender, age, location and schooling are the main markers of unemployment. They reject a 'jobless growth' explanation, because the employment elasticity of GDP rose to 0.80 in 1995-2005. Instead they point to rapid labour-force growth, skills-biased demand shifts that leave excess labour alongside a skills shortage, high union wage premia and relatively rigid labour regulation. They also stress the apartheid legacy of bifurcated schooling. The deck closes with policy options: reform higher education, improve job placement, improve regulatory efficiency, create a dual youth labour market and a youth cadet service, and pick winners in schooling.

Evidence and method

Post-1995 national household and labour force survey data, probably Statistics South Africa. The slides cite ILO (2007) and KILM 2007 for international comparisons, Fields et al. 1999 for wage-employment elasticities, and Statistics South Africa 1995 and 2000 with own calculations for the growth incidence curve. They also use a probit with a selection term (lambda), quantile wage regressions for union and bargaining-council premia, and cross-country labour regulation rigidity indices. No archival material is used.

Findings

  • Scale of unemployment in 2005: about 7.8 million people unemployed on the broad definition. (p. 5)

    Unemployed numbered approximately 7.8 million individuals in 2005

  • Discouraged work-seekers are a large share of broad unemployment, so the definition used matters for measuring unemployment. (p. 5)

    Discouraged work seekers constitute 42% of broadly unemployed

  • Race, gender, age, location and schooling are the key determinants of unemployment in the multivariate analysis. (p. 11)

    Race, Gender, Age, Location & Schooling serve as the key attributes defining and explaining broad and narrow unemployment rates in South Africa.

  • Apartheid schooling still shapes outcomes: disadvantage remains concentrated in former African schools, which show lower matric pass rates and higher pupil-teacher ratios. (p. 12)

    A bifurcated schooling system, where disadvantage is still located within former African Schools

  • Graduate unemployment is mainly a problem of diplomas and certificates, not degrees. (p. 13)

    Only 18% of unemployed graduates have degrees: Clearly the bulk of the problem is with diplomas and certificates

  • The 1995-2005 employment-GDP elasticity (0.80, against 0.16 in 1990-95) means that for the growth achieved, employment expansion was adequate. (p. 15)

    For Given Level of Growth, Employment Expansion Adequate

  • Labour demand shifted towards skilled workers (mostly within-sector shifts), producing excess labour supply alongside a skills shortage. (p. 19)

    The nature & trajectory of labour demand results in the co-existence of an excess supply of labour with a chronic skills shortage in the economy

  • Estimated wage-employment elasticities are in the normal range for developing countries (private sector total -0.35 in 1990-93 and -0.53 in 1994-98), which raises the question of wage adjustment. (p. 20)

    Elasticity value not unusual for developing country estimates

  • Regulatory rigidity is driven mainly by dismissal law, unfair labour practice provisions and fixed-term contract rules, though the indices measure legislation only. (p. 24)

    Source of Rigidity: legislation on dismissals; unfair labour practices; fixed-term contracts etc.

  • Growth in 1995-2000 favoured the non-poor, while state transfers after 2000 helped spread the gains. (p. 26)

    Growth path in the 1995-2000 period biased against the poor