Black Unemployment: A Case Study in a Peri-Urban Area of Natal¶

Literature › Works

Hofmeyr, J.F. (1984). Black Unemployment: A Case Study in a Peri-Urban Area of Natal. Second Carnegie Inquiry into Poverty and Development in Southern Africa, Carnegie Conference Paper No. 123. Cape Town (Second Carnegie Inquiry).


Type	conference paper
Scope	Economics and economic history of apartheid
Method	survey fieldwork
Period studied	1982–1983
Regions	Vulindlela (KwaZulu), Pietermaritzburg, Natal
Themes	inequality poverty unemployment, labour markets wages, bantustans removals influx, demography urbanisation, methods data sources
Citation key	`hofmeyr1984`

Argument¶

The paper asks whether Black male unemployment in the peri-urban Vulindlela area near Pietermaritzburg is mainly the result of too few jobs or partly a rational choice to search for better jobs. It sets a 'natural rate'/search model against a job-queue (job inadequacy) model, and tests them by comparing the employed, the unemployed and the voluntarily economically inactive on income, age, education, job tenure and job aspirations. The influx-control and labour-bureau system segments the market, most jobs are 'horizontal', and moving between jobs without a spell of unemployment is rare. Together these make holding out for a better job rational for those with household support. The evidence cannot decisively separate the two models. The author proposes a dual labour-market synthesis, with a high-wage sector whose wages are not set competitively and a low-wage competitive sector, and notes that the policy implications differ sharply.

Evidence and method¶

A household survey of 372 households (3,260 individuals), a roughly 10 per cent systematic sample of the Inadi I and II areas of Vulindlela surveyed by Bromberger in 1981. One working-age male (16-65) was selected at random per household. Fieldwork ran August 1982 to March 1983, using young Black matriculant interviewers and in-depth interviews of 1-3 hours. The paper also draws on the author's observation of the labour bureaux and on the secondary literature (Knight, Gerson, Maree, Bromberger, Harris and Sabot, Sen). Appendices cover definitions of unemployment, how the Pietermaritzburg labour bureau system works, and income distributions.

Place in the debates¶

Voluntary/search vs involuntary (job-shortage) explanations of Black unemployment in South Africa (see Unemployment: causes, measurement and labour market institutions). Inconclusive. The evidence is mostly consistent with both models, but the author favours a dual labour-market model. In it search or queuing for high-wage jobs explains much unemployment, while simple job inadequacy cannot be ruled out.
Demand-side structural unemployment (Maree) vs supply-side labour-market segmentation (Gerson, Knight) (see Unemployment: causes, measurement and labour market institutions). Sides with the segmentation and supply-side view for this peri-urban sample during an upswing or early recession, while stressing that the finding is specific to time and place.
Definition and measurement of unemployment (search, discouraged workers, underemployment) (see Unemployment: causes, measurement and labour market institutions). Adopts a recognition-based definition after Sen. A person must want work at the going wage and be actively searching. Low-income 'poverty' underemployment is excluded as a separate question.

Findings¶

Sample size and household structure of the Vulindlela survey. (p. 11)

The final sample consists of 372 households, comprising 3 260 individuals, giving a mean household size of 8,8.
The area is a low-cost commuter area, so low cash incomes go further than their size suggests. (p. 17)

Although incomes were low, the cost of living was also low,
Average unemployment spells are long (about a year), though most are short of three years. (p. 20)

The average duration of unemployment was 12,3 months.
Most jobs offer no promotion or training, so the only way to improve one's position is to change jobs. (p. 19)

It therefore appears clear that most jobs were effectively horizontal, or at least were perceived to be so by the incumbents
Most of the unemployed had household or other support that financed search. (p. 22)

Thus, it seems clear that most of the unemployed had some source of support apart from a regular wnge-paying job.
The labour bureau system segments the market into insider and outsider submarkets, with the best jobs going to those with urban rights. (p. 32)

In conclusion, perhaps the most important feature of the labour bureau system is that it segments the labour market into suhmarkets: one for urban insiders
Because jobs are horizontal and job-to-job moves are rare, holding out for better work yields high returns. (p. 32)

In other words, the rewards to time invested in job search are high.
Once the respondent's own wage is excluded, the employed come from poorer households than the unemployed or the economically inactive, which supports the idea that the unemployed could afford to search. (p. 40)

Once the wage income of the respondent is excluded, the households of the employed are significantly poorer than those of the other two categories.
Most of the unemployed would accept jobs at modest wages, which points to disequilibrium. They still refuse the least attractive jobs and any wage far below about R20. (p. 53)

fully 60 per cent were prepared .accept a job at R30 per week or less and 78 per cent at R35 per week or less.
The author concludes that a dual labour-market model fits the Black labour market better than a simple search or job-inadequacy model. (p. 66)

This suggests that a dual labour market type of model might describe the labour market for Blacks more accurately than either a simple search or a job-inadequacy model.
The job-inadequacy explanation cannot be ruled out, so the policy implications remain unresolved. (p. 68)

an explanation of unemployment based on simple job inadequacy cannot be ruled out by the evidence from the survey
Policy implication: remove labour-market segmentation and improve information flows rather than simply create jobs. (p. 70)

Improvement of information flows by eliminating the segmentation of the labour market caused by the labour bureau system would therefore be another way of reducing dispersion.

Evidence ledger¶

Sources used¶

Source	Kind	Period	How used
Author's household and in-depth male interview survey, Vulindlela (Inadi I and II areas), Pietermaritzburg, Aug 1982–Mar 1983	author survey or fieldwork	August 1982 – March 1983	Primary data source for household characteristics, employment, unemployment, labour-market behaviour, income comparisons and job aspirations of 372 households and selected working-age males.
Bromberger, 'Unemployment in South Africa: A Survey of Research' (Social Dynamics, 1978)	secondary literature compilation	1975-1977	Cited for gold mines experiencing breaking of contracts and labour shortage in 1975-76 despite high measured unemployment.
Industrial Council minimum wage for construction labourers (at start of survey, 90c per hour)	employer association records	c. August 1982	Used as the benchmark 'going market wage' for unskilled labour (R36/week) against which unemployed workseekers' reservation wages are compared.
Blacks (Urban Areas) Consolidation Act No. 25 of 1945, Section 10(1)(a),(b),(c)	legislation gazette	1945	Defines urban residence rights that segment the labour market into insiders and outsiders at the labour bureaux.

New data¶

Original survey of 372 households (3,260 individuals) in Vulindlela, Pietermaritzburg, a 10% systematic sample of ~4,500 households, with in-depth interviews of one randomly selected working-age male per household (about 353 selected males), covering employment histories, income, job search, labour bureau use and job aspirations; fieldwork Aug 1982–Mar 1983.

Inference¶

Quantification: Statistical analysis.

Single-site cross-sectional survey case study. Employed, unemployed and voluntarily inactive males are compared using cross-tabulations and chi-square-type tests of independence on household income, others working, reasons for job loss, education, age, job length and time out of work, to adjudicate between natural-rate/search and job-queue models. Reservation wages and job acceptance responses are used to test for market-clearing, leading to a dual labour market interpretation. Inference is inconclusive and limited by small sample size.

Key claims¶

Claim	Basis	Strength
Most unemployed males in the sample had other means of support (other household members or indefinite family support), so search unemployment was financially feasible for the large majority. (p. 18)	Author's survey: questions on sources of support and personal income of the 55 unemployed/underemployed.	Partial data
Most jobs held by the employed are effectively 'horizontal', with no promotion or training prospects, raising returns to holding out for a better job. (p. 15)	Survey questions on promotional prospects and training among the 215 full-time employed.	Partial data
Excluding the respondent's own wage, the households of the employed are significantly poorer than those of the unemployed or economically inactive, suggesting the unemployed could afford to be unemployed; but this is largely explained by age differences. (p. 36)	Survey household income cross-tabulations (Table 8.1, Appendix C) with pairwise significance tests, plus age controls.	Partial data

The author's caveats¶

Notes that the sample is somewhat biased against long-distance migrants; third-party wage reports are probably understated and informal income figures are unreliable; the age of older respondents is vague; small sample sizes meant expected cell frequencies were too small for many cross-tabulations, so duration of unemployment correlations were insignificant; the conclusions are specific to time (moving into recession) and place; some unemployed may have had unrecorded casual or gardening jobs; the inconsistent responses on mine job acceptance are acknowledged.

Gaps in the evidence¶

The study rests on a single small peri-urban sample (55 unemployed) with one cross-section, so it cannot test whether search behaviour varies across the business cycle or in truly rural areas, and it excludes women. Panel or longitudinal data following workseekers, vacancy and wage-offer data from labour bureaux and employers, and larger samples would sharpen the test between job-queue and search models. Administrative labour bureau registration and placement records existed at the time and could have been used to check reported job-search patterns.

Replicability: Data described but not public.