Global Poverty Research Group

Unemployment, race and poverty in South Africa

Overview

The unemployment rate in South Africa is one of the highest in the world, 36% to 42% since the year 2000 using the broad definition. Even according to the narrow definition, which applies a job-search test, 25-30% of adults who wanted work and actively looked for it were unemployed.  Moreover, the unemployment rates for different groups reveal great disparity in the incidence of unemployment. Given the importance of employment income in total household income in South Africa, the varying incidence of unemployment across different groups has important implications for the distribution of income and for the incidence of poverty. It is of policy interest to know how the incidence of unemployment varies by worker characteristics. That is, what attributes are associated with being able to avoid unemployment. Moreover, it is of particular interest, in the South African context, to examine to what extent differences in the productive and other observed characteristics of blacks and whites explain the race gap in unemployment. If the black-white differences in observed characteristics do not explain all the race-gap in unemployment rate, then this could be due either to labour market discrimination against blacks or due to the fact that some important characteristics that differ between blacks and whites are unobserved. For example, black and white labour force participants may have faced different school quality when they were of school-going age.

In this research, being undertaken by Geeta Kingdon and John Knight, the determinants of both entry into, and the duration of, unemployment in South Africa are being investigated using data from 1993, 1994 and 1997. The research aims to paint a picture of the distribution of unemployment in South Africa, asking the question ‘who are the unemployed?’ and identifying the characteristics that make a person more likely to be unemployed.

It is found that unemployment is very inequitably distributed in South Africa and certain groups are much more likely to enter it, and to stay in it, than others (Table below). Young uneducated Africans living in homelands and remote areas are most vulnerable to unemployment. There are two particularly striking features of South African unemployment: firstly, the fact that rural unemployment rates are higher than urban rates is atypical among countries and is explained by historical policies restricting mobility. Secondly, the majority (62%) of the unemployed have never held a job before, i.e., they entered unemployment from the time of entering the labour force. The very long duration of unemployment (>1 year) among a high proportion (68%) of the unemployed suggests that the demand-side of the labour market is responsible for a good part of the unemployment. 

Table 1

Unemployment rate (%), by age, education, gender, region, and race

 
Broad definition
Narrow definition
Broad-narrow gap
 
 
 
 
Race
 
 
 
African
41.2
26.2
15.0
Colored
23.3
19.4
3.9
Indian
17.1
14.3
2.8
White
6.3
4.2
2.1
 
Age
16-24
51.4
37.8
13.6
25-35
35.3
23.3
12.0
36-45
25.2
14.3
10.9
46-55
21.3
11.0
10.3
56-64
16.9
8.5
8.4
 
Education
 
 
 
None
38.7
20.1
18.6
Primary
42.5
26.8
15.7
Junior
35.3
23.5
11.8
Secondary
28.3
19.5
8.8
Higher
5.7
3.9
1.8
 
Gender
 
 
 
Male
26.2
17.3
8.9
Female
40.7
25.3
15.4
 
Region
 
 
 
Rural
40.3
23.4
16.9
Urban
27.9
19.1
8.8
 

Source: October Household Survey, 1994.

The analysis tells us the characteristics of the unfortunate people who are liable to be at the end of the queue for employment. Improving their characteristics may improve their place in the queue, but it will not necessarily reduce unemployment. In the African group - the group that suffers such catastrophically high unemployment rates - human capital characteristics such as education and employment experience dramatically reduce the chances of unemployment. However, a policy prescription that African education and skills should therefore be upgraded may not solve the problem: unless there are more jobs in the economy, upgrading the education of Africans will at best change the composition of employment in their favour. Of course, it is possible that expanding education and skills will reduce overall unemployment. The mechanism might be to increase the supply of skilled labour, for which there is market clearing, and to decrease the supply of unskilled labour, for which the market fails to clear and there is a surplus of workers.

The analysis suggests that racial differences in unemployment incidence cannot simply be dismissed as a problem of the poorer productive characteristics of the African, coloured, and Indian groups relative to the whites in South Africa. While a substantial part of the race gap in the incidence of unemployment in the mid-1990s was explained by inter-group differences in observed characteristics, there remained a large residual that could not be explained in this way. The residual may be due to employer discrimination or to racial differences in unmeasured determinants such as the quality of education. Further research incorporating data on the quality of education is being undertaken with available cross-section data, though longitudinal data sets are ideally needed to examine policy questions concerning unemployment dynamics.

Recent publications

Kingdon, G. and J. Knight, “Race and the incidence of unemployment in South Africa”, Review of Development Economics, 8, No. 3: 198-222. May 2004.

Kingdon, G. and J. Knight, “Unemployment in South Africa 1995-2003: Causes, Problems and Policies”, mimeo, GPRG, Department of Economics, University of Oxford, October, 2004.

Kingdon, G. and J. Knight, ‘What have we learnt about unemployment from microdatasets in South Africa?’, Social Dynamics, 27, No.1, 2002.

Kingdon, G. and J. Knight. “Quality of Schooling and the Race Gap in Labour Market Outcomes in South Africa”, mimeo, GPRG, Department of Economics, University of Oxford, March 2005.

Researchers to contact for this project

Geeta Kingdon & John Knight