Skip to content

Job Search in South Africa: A Nonparametric Analysis

Literature › Works

Martin Wittenberg (2002). Job Search in South Africa: A Nonparametric Analysis. The South African Journal of Economics, Vol. 70:8. Economic Society of South Africa.
Type journal article
Scope Post-1994 South African economy
Method descriptive quantitative
Period studied 1993–1995
Regions South Africa, urban vs rural areas
Themes inequality poverty unemployment, labour markets wages, education human capital, demography urbanisation, methods data sources
Citation key wittenberg2002

Argument

Wittenberg applies a search-theoretic, flows-based framework to South African unemployment, inferring flows from age-cohort stocks in three household surveys (1993 PSLSD, 1994 and 1995 OHS) using lowess nonparametric smoothing. He finds two distinct problems: a long-term labour absorption problem (only about 70 per cent of African men ever find employment) and an acute short-term unemployment 'spike' at about age 27 produced by late exit from school and slow absorption into work. The surveys are mutually consistent, so the very high unemployment levels are credible, and the falling male participation rate flagged by the ILO is largely a youth phenomenon. Education below matric makes little difference to employment prospects, which he reads as weak signalling that pushes employers toward insider networks and an insider-outsider division.

Evidence and method

Three cross-sectional household surveys (1993 PSLSD, October 1994 and October 1995 Household Surveys), analysed by age cohort, race, gender, urban/rural location and years of education. Also the ILO report on South Africa (Standing et al. 1996) including its employer recruitment-method data, Klasen and Woolard's 1998 re-analysis, and an anecdotal press report on 39,000 applicants to a University of Cape Town advertisement for gardeners and cleaners.

Findings

  • Two distinct unemployment problems: a long-term absorption failure and a youth spike around age 27. (p. 2)

    there is a long-term labour absorption problem: at maximum only 70 per cent of African men find employment. On the other hand, there is a very severe short-term unemployment ‘spike’ at around age 27.

  • African entry into employment is far slower than for Whites, and employment peaks at about 70 per cent. (p. 17)

    Whereas almost 100 per cent of all White men are in employment by age 25, this figure is still below 50 per cent in the case of Africans.

  • The strict unemployment peak is later for Africans, and peak levels differ by survey. (p. 17)

    In the 1995 OHS and the PSLSD it is at or just below 15 per cent, while in the 1994 OHS it is higher at about 20 per cent.

  • Long-term unemployment is dominant; many of the strictly unemployed have been jobless for years. (p. 21)

    nearly 80 per cent of the strictly unemployed at age 30 have been unemployed for at least a year.

  • A 'core' of roughly 10 per cent of the age cohort may never find a job despite searching. (p. 21)

    there is a ‘core’ of about 10 per cent of the age cohort that will essentially never find a job, but will still be undertaking active search activities

  • The ILO's declining-participation puzzle is a youth phenomenon rather than evidence of unmeasured activity. (p. 25)

    It is clear that this decline does not occur in the thirty-plus age groups. It is largely a youth phenomenon.

  • Urban residents show both higher employment and higher strict unemployment than rural residents. (p. 27)

    Urban people have both higher employment levels and higher unemployment levels!

  • Education below matric does not improve employment prospects; the effect comes only post-matric. (p. 29)

    the graph of the people working is essentially flat (perhaps even downwardly sloping) in all surveys until twelve years of schooling have been attained.

  • Informal recruitment via contacts dominates among firms (ILO data), supporting the insider-network interpretation. (p. 31)

    as their main method of recruitment, 41.4 per cent relied on friends and relatives of existing workers; 26.2 per cent used advertisements;

  • Discouragement appears to occur before personal job search experience, probably via information from social networks. (p. 24)

    many individuals seem to be discouraged before they have even attempted to find a job.