South Africa’s inequality problem is not only about how much the economy grows. New research suggests that the kinds of jobs provincial economies create, and the sectors in which people work, are closely tied to how evenly income is distributed.
A peer-reviewed study published on 26 September 2026 analysed all nine South African provinces over the period from the first quarter of 2008 to the fourth quarter of 2022. The researchers found that provinces with a greater concentration of employment in primary activities such as agriculture and mining tended to have higher inequality, while greater employment in secondary and tertiary activities was associated with lower inequality.
The result speaks directly to a long-running development debate. Moving workers from low-productivity activities into manufacturing and services can raise productivity and incomes, but economic transformation does not automatically distribute those gains evenly. For South Africa, the new analysis indicates that the structure of employment itself may be part of the inequality story.
The question behind structural transformation
Structural transformation describes the shift of economic activity and employment away from primary sectors and toward manufacturing, construction and services. Historically, this transition has accompanied economic development in many countries because workers and capital move toward activities capable of producing greater value.
Yet the transition can create a policy dilemma. The classic Kuznets argument proposes that inequality may rise during early development before later declining. New opportunities can initially benefit particular workers, firms or regions more than others. Governments may therefore face pressure to choose between reforms that promote transformation and policies aimed at limiting inequality.
Researchers Nosihle Cele, Syden Mishi and Robert Mwanyepedza of Nelson Mandela University asked how that dilemma plays out within South Africa. Their focus on provinces matters because national averages can hide sharply different regional economies. Gauteng is not structured like the Northern Cape, and the Western Cape does not have the same employment mix as Mpumalanga or Limpopo.
Nine provinces, 60 quarters and a changing economy
The study used regionally disaggregated Statistics South Africa data covering 2008Q1 through 2022Q4. That is a 15-year period comprising 60 quarters across nine provinces, giving the researchers a panel capable of tracking both differences between provinces and changes within them over time.
The analysis divided employment into broad economic sectors. Primary-sector employment captures activities such as agriculture and mining. Secondary employment includes activities associated with manufacturing and other forms of production, while tertiary employment covers the service economy.
The outcome of interest was inequality. Rather than asking only whether richer provinces were more equal or unequal, the researchers examined whether shifts in the employment composition of provincial economies were associated with changes in income distribution after accounting for the panel structure of the data.
The statistical challenge was substantial. Provinces do not evolve independently. National interest rates, commodity cycles, electricity constraints, recessions, fiscal policy and the COVID-19 shock can affect several provinces at once. The researchers detected cross-sectional dependence and slope heterogeneity in the provincial panel, meaning that conventional assumptions of independent regional observations were not appropriate.
They therefore used Driscoll-Kraay fixed-effects regression. This approach is designed to provide more reliable inference when panel observations can be correlated across units and over time. In practical terms, it allowed the study to focus on the relationship between changing provincial employment structures and inequality while using standard errors robust to important forms of dependence in the data.
Primary-sector concentration tracked higher inequality
The clearest pattern ran through primary-sector employment. Provinces with greater employment concentration in primary activities exhibited higher inequality. The result is important because mining and agriculture can generate substantial output without necessarily producing broad-based employment or distributing income evenly.
This does not mean that primary industries are inherently harmful or that South Africa should abandon sectors in which it has major natural and productive advantages. The more useful interpretation is that an economy heavily dependent on primary-sector jobs may struggle to translate resource production into a sufficiently broad ladder of incomes and employment opportunities.
The distinction between output and employment is central. A highly productive mine can contribute materially to provincial gross value added while employing a relatively small share of the population. Agriculture can likewise contain very different combinations of commercial productivity, seasonal work and low-paid labour. Economic value can therefore be substantial even when the employment structure remains unequal.
The study’s regional design strengthens this point. The researchers were not merely comparing South Africa with other countries that have different institutions and histories. They were examining variation inside the same national policy environment, across provinces exposed to common national institutions but different sectoral structures.
Manufacturing and services pointed in the opposite direction
Secondary and tertiary employment showed the opposite relationship. Expansion of employment in these parts of the economy was associated with more inclusive growth and a more equitable income distribution.
That finding supports the idea that structural transformation can become equalising when it creates a broader range of productive jobs. Manufacturing can connect raw materials to processing, logistics, engineering and supply chains. Services can range from retail and transport to finance, professional services, health, education and information-intensive work.
The result should not be interpreted as saying that every manufacturing or service job reduces inequality. South Africa’s service economy includes both highly paid professional work and precarious low-wage employment. Manufacturing also varies greatly in skill requirements, pay and labour intensity. The study instead identifies an aggregate provincial relationship: a larger employment role for secondary and tertiary activities was associated with lower inequality across the period analysed.
This is why the authors describe evidence consistent with a structural transformation dilemma. Economic development is not simply a matter of moving away from one sector and toward another. The distributional outcome depends on whether transformation generates employment that connects more people to productive parts of the economy.
Why value addition matters
The policy implications follow from the employment pattern. The researchers argue that South Africa should accelerate structural transformation rather than retreat from it, but should do so in ways designed to spread its gains.
One route is greater value addition in agriculture and mining. Instead of viewing a mineral or agricultural product mainly as an endpoint for extraction or cultivation, value addition asks how more stages of processing, manufacturing, logistics and associated services can occur within the economy.
That can matter for inequality because a longer domestic value chain potentially creates a wider set of occupations. The economic contribution of a commodity then extends beyond extraction toward processing, maintenance, technical services, transport, marketing and manufacturing. Whether this happens in practice depends on competitiveness, infrastructure, energy, skills and investment, but the employment logic is clear.
The authors also emphasise industry-relevant skills. Structural transformation can fail to become inclusive if expanding sectors demand capabilities that large parts of the workforce cannot access. Training policy therefore has a distributional role as well as a productivity role. The more closely education and skills development connect workers to growing sectors, the greater the chance that transformation broadens participation rather than merely raising output.
Special Economic Zones are part of the proposed toolkit
The study further points to Special Economic Zones as a mechanism for supporting industrial development. In principle, these zones can concentrate infrastructure, investment support and industrial activity in places where firms can build stronger production networks.
But the findings also imply a standard by which such policies should be judged. Attracting investment is not enough if new activity remains weakly connected to local employment and supplier networks. If structural transformation is expected to reduce inequality, the quality, accessibility and spread of employment matter alongside headline investment values.
This regional perspective is particularly relevant in South Africa because economic opportunity is geographically uneven. Place-based industrial policy can potentially help provinces develop productive specialisations, but a strategy built around local comparative advantage still needs connections to skills, transport, energy and markets.
What the study adds to the inequality debate
South African inequality is often discussed through wages, unemployment, education, race, wealth and redistribution. Those remain essential. The new research adds another layer: the sectoral composition of provincial employment.
This matters because redistribution operates after and alongside the economy’s initial distribution of opportunities and earnings. Structural transformation affects that earlier stage. If an economy creates more accessible productive employment, the labour market itself can distribute income more broadly before taxes and transfers are considered.
The analysis therefore connects industrial policy with inequality policy. Decisions about manufacturing capacity, value chains, skills and regional development are not only questions of GDP growth. They can also shape who has access to the kinds of jobs through which economic gains are distributed.
At the same time, the findings caution against assuming that economic modernisation automatically solves inequality. Transformation has to create employment at sufficient scale and across a range of skill levels. A province can become more technologically sophisticated without becoming more inclusive if the new economy excludes much of its workforce.
The analysis does not prove sector shifts cause inequality to change
The study has important limitations. Most fundamentally, it is an observational panel analysis. The regression identifies relationships between employment structure and inequality, but it cannot establish the same kind of causality that a randomised intervention could.
Reverse and simultaneous relationships are plausible. High inequality may itself influence the structure of provincial labour markets, investment patterns, migration and skills development. Other forces can affect both sectoral employment and inequality at the same time.
The broad sector categories also conceal substantial diversity. A tertiary job in a financial firm and one in low-wage personal services both sit within the service economy but have very different implications for income. The same applies to advanced manufacturing versus lower-productivity production, or capital-intensive mining versus labour-intensive agriculture.
The period from 2008 to 2022 includes unusually large shocks, including the global financial crisis, electricity constraints and the COVID-19 pandemic. The modelling strategy was selected partly because provinces are exposed to common shocks, but no statistical specification can turn such a turbulent period into a controlled experiment.
Finally, provincial averages can obscure inequality within cities, districts and communities. Structural transformation can occur unevenly inside a province, so future research at finer geographic levels could reveal where employment shifts are most strongly connected to inclusive outcomes.
Growth and distribution may be more connected than they appear
The study’s central message is not that South Africa must choose between growth and equality. It is that the composition of growth helps determine whether that trade-off becomes severe in the first place.
Across nine provinces and 60 quarters, greater dependence on primary-sector employment was associated with higher inequality, while stronger secondary and tertiary employment was associated with lower inequality. That pattern gives structural transformation a distinctly distributional dimension.
For policymakers, the implication is demanding but constructive. Value addition, industrial capacity, industry-relevant skills and place-based development can be assessed not only by how much output they generate, but by whether they expand the number and range of people able to participate in productive employment.
South Africa’s inequality challenge is too complex for any single policy lever. But this research suggests that changing what provincial economies do, and where their workers are employed, belongs inside the inequality conversation rather than outside it.
Source Information
Study: Socio-economic transformation and inequality in South Africa: a regional analysis
Authors: Nosihle Cele, Syden Mishi and Robert Mwanyepedza
Journal: Humanities and Social Sciences Communications
Published: 26 September 2026
DOI: 10.1057/s41599-026-09048-0
Study design: Provincial panel analysis using Driscoll-Kraay fixed-effects regression
Data: Statistics South Africa regionally disaggregated data for all nine provinces from 2008Q1 to 2022Q4
