Industrial specialisation, diversification, and concentration in post-apartheid Gauteng: 1994 - 2023

Introduction

Gauteng is South Africa's industrial heartland. While the province is also the country's commercial, financial and administrative centre it has historically held the most significant share of national mining and manufacturing production. Its economy evolved around what is today commonly referred to as the Minerals Energy Complex (MEC) (Fine & Rustomjee, 1996). Aside from mining and electricity, the MEC's core sectors included capital-intensive chemicals, iron and steel, non-ferrous metal industries, and other non-metallic mineral products.

Since 1994, Gauteng’s economy has faced considerable transformation as tertiary activities have increasingly gained prominence in its economic composition at the expense of its industrial activities. This transformation has been the result of increased international competition as South Africa’s economy opened up after apartheid era sanctions, structural issues inherited from apartheid, as well as government-led attempts to grapple with these. Since 1994, the provincial economy has seen a significant decline in the share contributed by formal agriculture, mining, and manufacturing – falling from 46% of total Gross Value Add (GVA)² to 31% in 2023. While experiencing relative decline, these industrial activities retain significance in economic development. This is supported by evidence of these activities’ positive contributions to growth and value creation (Mazzucato, 2019), manufacturing’s role as an “engine of growth” (Cantore et al., 2017), and substantial intersectoral linkages to the wider economy.

This GCRO Map of the Month builds on other recent maps that explore a variety of economic metrics across space (see Moisane & Murray, 2021; Naidoo, Maree & Naidoo, 2024; Hassen, Naidoo, Modiba, Mushongera, & Labuschagne, 2024). This month’s map investigates industrial specialisation (focusing on producing a specific type of goods or services), diversification (focusing on producing a wide variety of goods or services), and change in Gauteng’s municipalities and metropolitan planning regions between 1994 and 2023.

In this Map of the Month we use a measure of sector proficiency known as the Location Quotient (LQ)¹ for Gauteng’s municipalities and metropolitan planning regions to achieve two objectives. First, we look at which industry had the highest LQ in each area in 1994. This shows what the area was most specialised in at the time. We then compare this with the same picture in 2023 to show how each area's specialisation has changed. Second, we count the number of industries with LQs higher than 1 in 1994, and then in 2023 to show the degree to which there has been a change in the diversification or concentration of industrial activities.

LQ is interpreted as follows:

• LQ > 1: The geography (the planning region) has a higher share of the industry’s GVA, suggesting a specialisation or proficiency relative to the reference area (provincial or national economy).
• LQ = 1: The geography has the same share of the industry’s GVA, indicating no specialisation or proficiency relative to the reference area.
• LQ < 1: The geography has a lower share of the industry’s GVA, indicating a lack of proficiency or specialisation relative to the reference area.

In this map, the analysis is deliberately constrained to the agriculture, mining and manufacturing sectors, using the third level of Standard Industrial Classification (SIC) codes. Other sectors of the economy – including electricity, water & gas, and construction in the secondary sector, and all tertiary sector activities – are not considered. By excluding these other activities, the aim is to provide a clearer picture of the changes in Gauteng’s primary sector and manufacturing. However, this choice of focus does not negate the contributions that these other sectors make to the economy.

Gauteng's spatial industrial specialisation in 1994

In 1994, Gauteng’s economy was arranged around several industrial sectors that collectively accounted for 29% of the province’s total economic activity (measured by share of total GVA). Measured in constant 2015 prices, Gold at 7% was followed by Metals, Metal Products, Machinery & Equipment (5.2%); Food, Beverages & Tobacco (3.7%) and Petroleum products, chemicals, rubber and plastic (3.3%). This industrial composition was relatively stable over time, the result of the apartheid political economy’s creation of socio-spatial distortions and specialised industrial areas linked to mining (see Figure 1; Harrison & Zack, 2012; Ashman & Newman, 2018).

Figure 1. Spatial distribution of industrial specialisation (1994).

Notes: Specialisation is measured by the sector with the highest Location Quotient.

Disclaimer: The municipal boundaries have undergone significant spatial changes since 1994 (Mkhize & Khanyile, 2020). Because the data used in this Map of the Month is sourced from Quantec, the regional mapping is based on the 2016 municipal demarcations.

Figure 1 shows the economic sector with the highest Location Quotient in each Gauteng municipality and metropolitan planning region in 1994. For example, in 1994, Lesedi’s highest LQ was in Coal (QSIC 21) at 33.82, indicating its significant contribution to Gauteng’s coal industry share of the province’s total economic value. A crucial point to note here is that “specialisation” as it is used in this Map of the Month does not imply a region was not proficient in other industrial activities. Along with coal, Lesedi was also proficient in Food, Beverages & Tobacco (LQ = 6.63), Agriculture (LQ = 6.26) and Metals (LQ = 1.73).

Low-value-adding, labour-intensive industries such as agriculture and fishing (shades of green) and textiles (light purple) were concentrated in peripheral regions and townships, leveraging displaced black labour from Bantustans and disadvantaged urban areas like Soweto (Mosiane & Gotz, 2022). Capital–intensive manufacturing activities (darker shades of purple), known for higher value addition, wages, and productivity (Cantore et al., 2017) were located in closer proximity to white urban areas and agglomerations of mining activity (shares of orange). These included Midvaal, Lesedi, and Emfuleni in the Sedibeng district, parts of the City of Johannesburg (JHB-G, JHB-F) and portions of the City of Ekurhuleni (EKU-A, EKU-D, EKU-E).

Apartheid-era spatial-industrial planning is also evident in Emfuleni and EKU-F (Alberton to Vosloorus) where proficiency in Metals, Metal Products, and Machinery & Equipment (QSIC 35) activities were supported. Specialising in these activities was done with the aim of fostering value-adding linkages to neighbouring mining-specialised areas. Other areas of manufacturing specialisations close to mining inputs include Transport Equipment (QSIC 38) in parts of Tshwane (TSH-1, TSH-6) and EKU-C (Daveyton).

Changes to Gauteng’s industrial specialisations across its geography between 1994 and 2023

In the post-apartheid period, Gauteng’s space economy underwent a significant restructuring. This was partly due to structural factors and partly due to the government's efforts to restructure the province’s uneven industrial landscape, redressing apartheid’s spatial legacies and managing the country’s response to globalisation. By 2023, the spatial distribution of industrial specialisations shows a retreat from higher-value manufacturing sectors (darker shades of purple) to lower-value primary sectors (shades of green and orange) across the province (Figure 2).

Figure 2. Spatial distribution of industrial specialisation (2023).

Notes: Specialisation is measured by the sector with the highest Location Quotient.

Emfuleni and EKU-F (Alberton to Vosloorus) provide examples. Previously specialising in metals, metal products, machinery, and equipment (QSIC 35), they now appear to prioritise other mining and quarrying (QSIC 22, 25-29) and fishing (QSIC 13), respectively. For example, Emfuleni’s metals, metal products, machinery, and equipment sector accounts for a much smaller share of its total economic activity (8% in 2023 compared to 15% in 1994). Similar declines occurred in EKU-F, where the share of its total economic activity accounted for by its metals, metal products, machinery, and equipment sector declined from 8% in 1994 to 3% in 2023.

The data also shows that Tshwane’s east region (TSH-6), previously specialised in transport equipment (QSIC 38), has become relatively more proficient in forestry (QSIC 12). However, a deeper look at GVA trends between 1994 and 2023 reveals stark differences in economic contribution. While TSH-6 became relatively proficient in forestry compared to its other economic activities, this sector averaged less than 1% of the region’s total economic GVA. Meanwhile, TSH-6’s transport equipment’s share declined slightly from 2% in 1994 to 1% in 2023.

While the above analysis provides a high-level overview of changes in the Gauteng province’s industrial structure between 1994 and 2023, it does have some limitations. It is important to note that some apparent shifts in production proficiency may be due to the nature of the LQ measure itself, rather than actual structural changes in economic activity. For example, forestry’s share of GVA in Gauteng has declined relative to the transport equipment sub-sector’s growth, affecting the calculation. Additionally, this analysis does not account for shifts in other sectors such as electricity, gas, water, construction, and tertiary sectors or the broader impact of urbanisation patterns on industrial composition over time.

Critically, while the analysis does provide high-level, static insights into how the province’s industrial sector composition has changed over time, it thus far lacks nuance on the changes to the basket of industrial activities being produced in different planning regions. The dynamics implied by the changing landscape of industrial specialisation speak to the need for deeper investigations into the composition of the scope and scale of industrial activities, how these have changed over time, and how they are impacted by other changes in the composition of economic activity in these planning regions. In service of these efforts, the next section explores the scope and scale of diversified industrial activities across the Gauteng geography.

Spatial diversification and concentration in the post-apartheid Gauteng industrial economy, 1994 and 2023

While the previous section discussed specialisation and proficiency, it was cautious to note that none of these regions were proficient (i.e. possessed an LQ > 1) in only one set of industrial activities. Instead, the data shows that in 1994 and 2023, many planning regions were proficient in multiple industrial sectors (refer to Figure 3). Possessing proficiency in multiple industrial activities is termed industrial diversification. Conversely, if a region possesses few proficiencies, it can be thought of as being industrially concentrated. According to economic theory, diversification enhances resilience to industrial shocks, whereas regions exhibiting concentration around a few industries would be more at risk and susceptible to industry-specific shocks.

The data indicates that in 1994 Ekurhuleni was the most diversified municipality, with all its planning regions containing 10 or more sub-sectors. Similarly, Johannesburg exhibited relatively strong levels of diversification in 1994, with regions ranging from four sectors (JHB-B, Randburg) to 10 (JHB-A, Midrand; JHB-G, Orange Farm). In contrast, Tshwane had lower numbers of sectors with LQ > 1 (i.e. was relatively concentrated around a few industrial activities), averaging 4.9 sectors across its seven planning regions. TSH-7 (Bronkhorstspruit) was the most diversified (9 sectors), followed by TSH-2 (Sinoville/Hammanskraal) with six.

Figure 3 speaks to the changes in diversification and concentration between 1994 and 2023 (the latest available data). In this regard, the Sedibeng district (Emfuleni, Midvaal, and Lesedi) have all increased the number of sectors in which they were proficient (adding 7, 4, and 8 sectors, respectively). Thus, in terms of our analysis, these regions became more diversified. Similarly, Mogale City, Merafong, and Rand West have increased their degree of industrial diversification to 14, 7, and 9 sectors from 10, 2, and 4, respectively. Midrand (JHB-A) stands out in terms of the extent of its decline in industrial sectors in which it was proficient over the period, becoming more concentrated since 1994.

Figure 3. Changes in industrial diversification and concentration in Gauteng's municipalities and metropolitan planning regions, 1994 and 2023.

Note: Diversification and concentration is measured by the number of sectors with Location Quotients greater than 1. Pink lines indicate declines in the number of diversified industrial activities (i.e. concentration). Blue lines indicate increases in a region’s diversification (i.e. more industrial activities with LQ > 1).

Spatially, the historical industrial cores of Ekurhuleni and Johannesburg have, in all but one region (Kempton Park to Boksburg), become more concentrated (indicated by darker shades of pink) to varying degrees (Figure 4). In contrast to these declining historical industrial centres, Tshwane has been a bright spot since 1994. All of its planning regions, except Centurion, either became more diversified (indicated by darker shades of blue) or remained unchanged (indicated by grey) since 1994.

Figure 4. Differences in industrial diversification and concentration in Gauteng's municipalities and metropolitan planning regions, 1994 and 2023.

Notes: Diversification is measured by sectors with Location Quotients greater than 1 and its intensity is indicated by darker shades of blue. In contrast, areas that have become more concentrated are denoted by darker shades of pink.

It is important to note that despite Tshwane’s economy becoming more diversified since 1994, its share of Gauteng’s total GVA has remained unchanged at 18%. This is noteworthy in comparison to other municipalities where, for example, the share of industrial diversification in Johannesburg and Ekurhuleni decreased significantly. Specifically, Johannesburg saw a drop in its industrial GVA share from 28% to 18% of its total GVA, while Ekurhuleni’s share decreased from 36% to 31%. Johannesburg’s industrial decline can be seen through the commensurate growth and dominance of the tertiary sector in the post-apartheid period (specifically the FIRE sectors; see also Gotz & Todes, 2014).

Other areas, particularly Sedibeng (Lesedi, Midvaal, and Emfuleni) and West Rand (Merafong, Rand West, and Mogale City), showed substantial gains in the number of industrial activities in which they became proficient since 1994. The rise in these areas’ share of Gauteng’s diversified industrial activity, in the case of Sedibeng from 9% in 1994 to 18% in 1993, and in the case of the West Rand from 9% to 15%, aligns with this.

Looking more closely at the Sedibeng District Municipality, Emfuleni, Midvaal, and Lesedi have experienced contrasting changes in the share of their total GVA contributed by industrial sectors. For instance, between 1994 and 2023, Emfuleni’s industrial sectors’ combined shares of total economic GVA declined from 31.7% to 28.6%. The sub-sector behind this decline was metals, metal products, machinery and equipment [QSIC 35]. Despite this decline, Emfuleni still maintained a robust LQ with its metals, metal products, machinery and equipment sector ranking second among its seven sectors in 2023(LQ = 2.6; down from 3 in 1994).

Implications and conclusions

This Map of the Month highlights shifts in Gauteng’s structure and composition of industrial activities between 1994 and 2023. It offers insights into the changes in industrial proficiency across the province’s geography and the degree to which different planning regions have become more diversified or concentrated. However, the analysis suggests that further investigation is needed to understand standout planning regions. For instance, Mogale City and other regions within the Sedibeng and West Rand District Municipalities warrant closer examination when compared to the changes that have occurred in the historical industrial centres of Johannesburg and Ekurhuleni.

The analysis provides a contribution to two areas of research. First, by analysing the changes to the degree of diversification and concentration, the extent of linkages within and across planning regions, and their sub-sectoral interactions, this Map provides valuable insights that can inform and shape future provincial and regional industrial policies. This includes examining how regional product spaces³ interact with broader provincial industrial and urban economies and local value chain activities permeating across the province. Integrating spatial considerations into industrial development strategies will enable Gauteng’s policymakers to design targeted policies that address regional needs while fostering synergies across economic, social, and urban dimensions.

Second, the Map is also well-positioned as an entry point into the academic literature surrounding spatial distribution and industrial agglomeration economies. It provides a foundation upon which to explore the dynamics of industrial diversification and concentration in relation to the changing structure of economic activities within and across regional geographies.

While this analysis offers valuable insights for policymakers and academics by capturing geographical structural changes in industrial activities over time, subsequent analyses can strengthen the insights in several areas. The analysis could be complemented with deeper investigations into specific industrial sectoral changes, focusing on product spaces, industrial linkages, and evolving global, regional, and local value chains. Further analysis would also need to locate these industrialisation dynamics in relation to the non-industrial sectors of the economy that were deliberately excluded at the start.

Footnotes on method and metrics

¹ Location quotients are used to identify sectors that are clustered in a given region or geography where a region has a proficiency in a given economic activity compared to a reference geography (Porter, 1990; O’Donoghue & Gleave, 2004; Carroll et al., 2008; Tian, 2013). However, the Location Quotient is not without its limitations. These are discussed in detail in Pominova et al. (2021) and recognised in this Map of the Month. Moreover, the construction of the LQ metric is sensitive to industry aggregations, regional definitions, and benchmark selection (Pominova et al., 2021). Because of this, the chosen construction of LQ can produce different interpretations of a geography’s specialisations and diversification or concentration.

To calculate the Location Quotient:

² Gross Value Added (GVA) is an economic metric that measures the total value of all goods and services at the price at which these are produced in an economy. It is a component of GDP, with GDP being the sum of GVA for all economic sectors, adjusted for taxes and subsidies. All data was sourced from Quantec. Dataset: Location Quotient for GVA at basic prices by 2016 municipality/ward-based metro region (Available here).

³ The product space maps the relationships between different products based on their underlying productive capabilities and the knowledge required to produce them. It offers insights into why some geographies are better positioned to diversify into certain industries.

References

Ashman, S. and Newman, S., 2018. The Evolution of Manufacturing in the Gauteng City-Region: From De-Industrialization to Re-Industrialization?. The Changing Space Economy of City-Regions: The Gauteng City-Region, South Africa, pp.131-156.

Cantore, N., Clara, M., Lavopa, A. and Soare, C., 2017. Manufacturing as an engine of growth: Which is the best fuel? Structural Change and Economic Dynamics, 42, pp.56-66.

Fine, B. and Rustomjee, Z., 1996. The Political Economy of South Africa: From Minerals Energy Complex to Industrialisation, Boulder, CO: Westview Press.

Gotz, G. and Todes, A., 2014. Johannesburg’s urban space economy. Changing space, changing city: Johannesburg after apartheid, pp.117-136.

Harrison, P. and Zack, T., 2012. The power of mining: the fall of gold and rise of Johannesburg. Journal of Contemporary African Studies, 30(4), pp.551-570.

Hassen, E.-K., Naidoo, L., Modiba, M., Mushongera, D., and Labuschagne, H., 2024. Spatial changes in Gauteng’s formal manufacturing jobs: 2014 – 2022. GCRO Map of the Month. DOI: https://doi.org/10.36634/LDPH9021.

Mazzucato, M., 2019. The value of everything. Making and taking in the global economy. Penguin London, UK: Random House UK.

Mkhize, T. and Khanyile, S., 2020. The changing municipal and provincial boundaries of Gauteng. GCRO Map of the Month (January 2020). DOI: https://doi.org/10.36634/BSYN8841

Moisane, N. and Murray, J., 2021. Distribution of population vs economic activity across the GCR. GCRO Map of the Month. DOI: https://doi.org/10.36634/OQIS4226

Mubiwa, B. and Annegarn, H., 2013. Historical spatial change in the Gauteng City-Region. Available here: https://d3iy114y1zl2zq.cloudfront.net/media/documents/mubiwe_occasional_paper_new.pdf

Naidoo, Y., Maree, G. and Naidoo, L.., 2024. A method for mapping high resolution Gross Value Added and its changes between 2012 and 2021 in Gauteng. GCRO Map of the Month. DOI: https://doi.org/10.36634/ENXB6242

Pominova, M., Gabe, T.M. and Crawley, A., 2021. The pitfalls of using location quotients to identify clusters and represent industry specialization in small regions. International Finance Discussion Paper, (1329).

Inputs, edits, and comments: Graeme Götz, Dr Laven Naidoo, Dr Samkelisiwe Khanyile, and Christian Hamann

Suggested citation: Bell, J.F. (2025). Industrial Specialisation, Diversification, and Concentration in Post-Apartheid Gauteng: 1994 - 2023. GCRO Map of the Month, Gauteng City-Region Observatory, March 2025.