Internal Migration of Blacks in South Africa: Self-Selection and Brain Drain

Literature › Works

Chung Choe; E. LaBrent Chrite (2006). Internal Migration of Blacks in South Africa: Self-Selection and Brain Drain. University of Arizona (draft; first draft Nov 2006, this draft Nov 2007).

Type	working paper
Scope	Post-1994 South African economy
Method	econometric
Period studied	1992–1996
Regions	South Africa, 318 magisterial districts, nine provinces
Themes	demography urbanisation, labour markets wages, education human capital, transition post apartheid, bantustans removals influx, methods data sources
Citation key	choe2006

Argument

The paper treats the dismantling of apartheid movement controls as a quasi-experiment in how skilled and unskilled black workers choose locations. Using the 1996 census and a conditional logit over 318 magisterial districts, it finds that black movers sought higher expected log wages, as Roy and Sjaastad predict. Movers with some post-matriculation schooling favoured districts with a higher share of high-schooled residents, while less-educated movers avoided them. The authors conclude that internal migration produced a within-country 'brain drain' and greater segregation along education lines. Returns to unobservable skill are insignificant or wrongly signed.

Evidence and method

1996 South African Population Census (10% unit-level sample), with previous-residence information for 1992-96 movers: 34,778 movers and 162,542 stayers among black male paid employees aged 18-62. Income is coded in 14 categories and converted to midpoints, with a Pareto-curve value for the open-ended top category. Also 1996 SAPS crime data by precinct, GIS-derived district centroid distances, and 1991 vs 1996 census shares of the population with post-matriculation schooling.

Findings

• Apartheid was backed by a large body of restrictive law, with the final migration-relevant controls repealed in 1991. (p. 2)

  Between 1947 and 1991 the South African government had established an Apartheid system with up to 317 laws that discriminated against black workers.

• The pass laws tied a black person's right to be in an area to approved work in a single magisterial district and allowed 'endorsing out'. (p. 6)

  A pass was issued for one ‘magisterial district’ (usually one town) confining the holder to that area only.

• The pass laws produced mass arrests. (p. 7)

  Each year, over 250,000 blacks were arrested for technical offenses under the Pass Laws.

• Black educational access was severely restricted under Bantu Education. (p. 7)

  Only 14 percent of black students reached secondary schools, and even fewer furthered their education to the university level.

• Black socioeconomic status remained far below white in 1996, with high black unemployment. (p. 16)

  The mean income levels of whites were roughly three times more than those of blacks.

• Black movers were younger, better educated and more often employed than stayers, in a labour market with 36% black unemployment. (p. 16)

  The unemployment rate of blacks is 36% for the entire sample and movers are more likely to be employed

• Among less-educated workers, the wage differential attracts migrants while districts with a higher share of high-schooled residents repel them. (p. 24)

  the coefficient of the share of high schooling in area j minus the share of high schooling in origin is significantly negative consistently across various specifications.

• Workers with some matriculation are attracted to districts with a higher share of educated residents, producing a reshuffling of educated people. (p. 24)

  they tended to be attracted to areas where a higher percentage of people had some high school education.

• The return to unobservable skill is insignificant for the less-educated and unexpectedly negative for the more-educated, contrary to the Roy prediction. (p. 24)

  the coefficient on the returns to unobservable characteristics is unexpectedly negative.

• The authors conclude that migration after apartheid increased segregation by education. (p. 25)

  Thus, the migration process contributed to greater segregation along education lines.