Our previous analysis on graduate unemployment found that young black female graduates face significantly higher unemployment rates compared to other segments of graduates. It is estimated that unemployment rates for young black females are nearly ten times higher than their white male counterparts. One possible explanation is that the labour market rewards some universities and fields of study more strongly than others, resulting in uneven employment outcomes across graduates.

The analysis also highlighted the need for better data to understand these dynamics more clearly. In particular, linking SARS tax records with data from the Higher Education Management Information System (HEMIS) could significantly improve our understanding of graduate employment trajectories. Since both administrative datasets use ID numbers as unique identifiers, these records could be securely linked using privacy-protecting methods such as hashing and salting. This would still allow researchers to track whether graduates enter formal employment, how long it takes them to do so after graduation, and what they earn once employed.

We unfortunately do not have access to SARS tax data (yet). But, what we can do is explore graduate supply trends using HEMIS data, specifically who is graduating, from where, and in what fields.

More graduates, and a very different profile

According to HEMIS data, the number of students graduating from South African universities increased from 104,744 in 2008 to 179,625 in 2023 (the latest year of publicly available data). That is a compound annual growth rate of 3.7%, well ahead of annual population growth of 1.4%.

This growth has not been demographically neutral. Female graduates have grown at 4.3% per year over the period, compared to 2.5% for male graduates. By 2023, two out of every three graduates were female.

Over half of total graduate growth between 2008 and 2023 comes from a single institution – UNISA. Graduates from UNISA more than tripled over the period, from under 18,000 to over 56,500. In 2008, less than one in five graduates came from UNISA. By 2023, it had increased to almost one in three.

The racial composition of graduates has also shifted markedly. In 2008, white males accounted for 14% of all graduates or 14,744 in absolute terms. By 2023, that had fallen to under 10,000, or 3.5% of all graduates. Over the same period, black female graduates grew from 35,000 in 2008 (34% of graduates) to over 91,000 in 2023 (51%) – the fastest growing segment by a considerable margin.

The top four race/ gender segments by number of graduates are shown in the chart below. Explore the data further by filtering on a specific university.

UNISA alone accounts for more than half of the increase in black female graduates over the period. This has important implications for how we interpret patterns of graduate unemployment.

While there is no direct evidence (yet) that UNISA graduates experience systematically poorer employment outcomes than graduates from contact universities, the possibility is plausible and deserves closer scrutiny. Distance-learning students may face disadvantages in employer signalling, access to alumni and peer networks, career services, and the forms of informal professional socialisation that campus life often provides; the absence of an in-person university experience may affect the development of social capital and the cultural fluency needed to navigate professional environments with confidence.

This is ultimately an empirical question. A linkage between HEMIS and SARS administrative data would make it possible to rigorously test whether labour market outcomes differ systematically between graduates from distance and contact institutions.

What degrees do graduates have?

The field of study also varies notably by demographic group. Education is the most popular field of study for black females, with over 23,000 graduating in 2023, again, over half of them from UNISA. Black female graduates are also over-represented in Social Sciences and Public Management, and significantly under-represented in Computer Science and Engineering, and to a lesser extent in Health Professions.

Fields like Engineering and Computer Science may be absorbed more readily by the labour market, while Education and Social Sciences are more dependent on public sector hiring, which is often constrained and slow-moving . The concentration of black female graduates in these fields may partially explain the unemployment gap identified in the previous analysis.

Filter on different demographic segments in the chart below:

The trend in field of study is also interesting. In 2018, Education was the most popular field, with almost 48,000 graduates. By 2023, that had dropped to just under 37,000 – a decline of over 11,000 graduates in five years. The drop is driven largely by UNISA, which produced 8,750 fewer Education graduates in 2023 compared to 2018. What drove this shift is not clear. Possibilities include a contraction in public sector demand for teachers, changes in UNISA’s programmes offerings, or shifting student preferences.

What the HEMIS data can and cannot tell us

That analysis of HEMIS data points to two structural features worth noting. First, a single distance-based learning institution has a significant impact on the profile of South African graduate output. UNISA has played a vital role in expanding access to higher education for South Africans who may not otherwise have attended university with almost one in three graduates coming out of UNISA. Second, there are noticeable differences in fields of study across demographic segments. While HEMIS data helps us frame the critical question, it alone cannot answer it: which graduates find employment, and which do not? Linking HEMIS to SARS administrative data would make this question answerable. The resulting evidence would meaningfully strengthen the decisions made by students choosing institutions, universities designing programmes, and policymakers allocating resources.