Skip to content

Export Diversity and Regional Growth: Empirical Evidence from South Africa

Literature › Works

Marianne Matthee; Wim Naudé (2007). Export Diversity and Regional Growth: Empirical Evidence from South Africa. UNU-WIDER Research Paper No. 2007/11. UNU-WIDER, Helsinki.
Type working paper
Scope Post-1994 South African economy
Method econometric
Period studied 1996–2004
Regions South Africa, 354 magisterial districts, Gauteng, KwaZulu-Natal, Eastern Cape, Western Cape
Themes manufacturing industrial policy, comparative theory, methods data sources, transition post apartheid
Citation key matthee2007

Argument

The paper asks whether what a region exports, and not just how much, matters for subnational growth in post-apartheid South Africa. It builds diversity measures (Herfindahl, Al-Marhubi, normalised Hirschmann, and the PRODY/EXPY indices of Hausmann et al.) from export data for 19 sectors in 354 magisterial districts. Barro-type OLS regressions find that more diversified exporting districts grew faster, and cubic-spline plots suggest diversity falls with distance from ports. It reads this through new economic geography and suggests FDI may be pulling diverse export activity toward ports.

Evidence and method

Subnational export data for 19 industries from the South African Revenue Services (Customs and Excise), 1996-2004. GDP, population and education data come from Global Insight's Regional Economic Explorer, built on Statistics South Africa sources. Distances to the four export hubs come from Shell Geostar. The authors' own calculations underlie the maps, cubic-spline plots and OLS regressions on 281 exporting districts.

Findings

  • Districts with highly concentrated exports grew slowly and contributed little to exports. (p. 8)

    The magisterial districts with an index value greater than 0.90 in 2004 experienced an average annual real GDP per capita growth rate below the average for all exporting magisterial districts in 2004.

  • Highly diversified districts supplied about a third of South Africa's exports in 2004. (p. 8)

    The contribution made to total exports in 2004 is 32.90 per cent.

  • Diversified districts export in most sectors, concentrated districts in very few. (p. 8)

    the more diversified districts produce exports in 17 of the 19 sectors, whereas the more concentrated districts produce exports only in 3 sectors

  • Diverse-export districts cluster near ports. (p. 12)

    It appears that those magisterial districts with a diverse range of exports are located within around 100 km from the nearest export hub.

  • Concentration indices are significantly and negatively related to GDP growth, while controls (human capital, openness, distance, population growth) are insignificant. (p. 14)

    The negativity implies that, with other given factors, larger export diversification and lower concentration or specialization contributes to real GDP growth.

  • Distance has a negative but insignificant coefficient in the growth regressions. (p. 14)

    The coefficient on distance, which proxies for domestic transport costs, is negative in all three instances, although not significant.

  • Vertical diversification away from primary exports is not a significant growth driver locally; horizontal diversification in manufacturing is. (p. 15)

    vertical integration in South Africa is not a significant source of economic growth on local level. Horizontal diversification (in manufacturing), however, is associated with larger growth

  • Distance has become more important since 1996, with fewer diverse districts far from ports by 2004; FDI is offered as a tentative explanation. (p. 16)

    distance (transport costs) has become more important since 1996 (under greater openness), with fewer diverse magisterial districts located further away from ports in 2004 than in 1996.