Moving early childhood development to Basic Education will not fix subsidy design, practitioner pay or access
Introduction
When responsibility for early childhood development (ECD) moved from the Department of Social Development to the Department of Basic Education in 2022, it was widely welcomed as a positive step (Matlala and Molokwane, 2024). Announced in 2019, the shift fits a wider pattern of treating the years before school as an education matter and not only a child protection one (Karisa et al., 2022). The hope is reasonable. This paper argues, though, that the shift on its own leaves untouched the things that decide whether poor children get a decent early learning programme: how the subsidy is designed, what practitioners are paid and trained for, and whether centres in poor communities can survive on fees their parents can barely afford.
The discussion has four parts. The first asks what the function shift actually moves and what it leaves behind. The second looks at registration and the subsidy, where money follows paperwork (Sello, 2025; Sello et al., 2024). The third turns to practitioners, who are the people the whole system depends on (Zulu et al., 2022; Adendorff and Bipath, 2025). The fourth considers quality and access in poor communities, drawing on a mix of census analysis, simulation and small case studies (Venter, 2022; Moses and van der Berg, 2023). Much of the evidence comes from Gauteng or from single census snapshots, so the claims below are made with that in mind.
What the function shift moves, and what it leaves behind
The shift is narrower than the headline suggests. Karisa et al. (2022) set out what goes to the education department: the early learning curriculum, continuity with Grade R, training and monitoring of curriculum delivery, and the registration and quality monitoring of ECD programmes. Child protection, psychosocial services, the child support grant and partial care stay with Social Development. The rationale is that ECD should be led as an early learning function, in line with a global trend. Yet the same authors list the questions stakeholders were already asking, such as what happens to existing centres and staff and how funding for new centres will be found (Karisa et al., 2022; Matlala and Molokwane, 2024).
Matlala and Molokwane (2024) went to the practitioners themselves. They interviewed staff in ten Gauteng centres, six unregistered and four registered, in Tembisa, Soweto, Soshanguve, Capital Park and Gezina, about what the move might do to their jobs. Reactions were mixed. Some welcomed the prospect of regular income, health benefits and a department that might answer their calls, while others doubted that the new department could absorb everyone and complained about not being consulted. Six of the ten were unhappy with how little the government recognised their self-taught skills and experience. The authors point out that the government had given no clear projection of how many practitioners the education department would need or what it would require of them (Matlala and Molokwane, 2024).
That uncertainty may not be an accident. Adendorff and Bipath (2025) report that the education department has said it will not become the employer of ECD practitioners, and that its current focus is getting everyone to at least the entry-level Level 4 qualification. They conclude that, in the medium term, the new higher education qualifications policy is unlikely to change much for people working with children from birth to four. Put next to the optimism around the shift, this is a sobering finding. A change in the department’s name on the letterhead is not a change in who pays the practitioner, and that is the question that matters to most of the workforce (Adendorff and Bipath, 2025; Matlala and Molokwane, 2024).
The scale of what is being moved is worth stating plainly. The 2021 census counted 42,420 early learning programmes and 1.6 million children, and 93.7 per cent of programmes charged fees (Sello, 2025). Adendorff and Bipath (2025) describe centres as small businesses with unstable finances, many of them depending on fees and struggling with registration. So the department inherits a sector that is mostly privately run and fee-funded, with the state holding few levers beyond the subsidy and the registration rules. That matters because school-style governance assumes the state is the provider, which is not the case here.
Registration and the subsidy: money follows paperwork
The subsidy is the main lever the state holds, and the census evidence shows how tightly it is tied to formal status. Sello (2025) analysed the 2021 ECD census, using data from 39,176 centres that answered the registration question. About 68 per cent were registered as non-profit organisations, and receiving a subsidy was the strongest predictor of registration, with an odds ratio of 10.17. Only around 33 per cent of centres received the Social Development subsidy at all. The author reads this as evidence that subsidies push centres to formalise, though the data are cross-sectional and cannot settle which comes first (Sello, 2025).
Registration itself is not cheap or simple. Sello et al. (2024) analysed 39,375 centres and found 41 per cent registered, 43 per cent not and 16 per cent in the process. Lacking a bank account was the strongest predictor of registration barriers, and among unregistered centres about three-quarters had no account. Roughly 60 per cent of unregistered centres put all children in one room, and almost 70 per cent had one or two staff. Registration can require a land use certificate, building plans, and fire, health and food clearances, while the department funds neither renovations nor title deeds (Sello et al., 2024). In other words, the centres least able to meet the rules are the ones in the poorest places, and the subsidy sits behind those rules.
What the subsidy pays is another matter. Sello (2025) notes that the rate of R17 per child per day stayed fixed from 2019 until it rose to R24 in March 2025. One practitioner told Adendorff and Bipath (2025) that the subsidy did not cover electricity, gas and food, and centres across their sample lean on parent fees to cover salaries and utilities. The education department also has no costed plan for its qualifications policy. In the Durban study by Zulu et al. (2022), the unregistered centre could not claim the subsidy at all, charged R250 a month and paid its practitioners about R800. The picture suggests a funding model that rewards formality and undervalues the poorest providers.
There are two ways to read these numbers. One says the subsidy works as an incentive and should be kept as a reward for formalising, which is roughly where Sello (2025) lands while also asking for targeted help with infrastructure and administration. The other, closer to what Matlala and Molokwane (2024) heard, says unregistered operators mostly want to comply but find the process too complicated, especially in poor communities. The two are not far apart. Both suggest that the barrier is capacity, not willingness, and that a subsidy which pays only after a centre has cleared every hurdle will keep missing the centres that need it most (Sello et al., 2024).
Registration rules also changed shape in the middle of this. The education department dropped the requirement that centres register as non-profit organisations in January 2023, intended to cut paperwork. Sello (2025) welcomes the aim but warns that some may read the change as meaning no registration is needed at all, which could push practitioners towards operating illegally and put future funding at risk, since the Children’s Act still requires registration. The census predates the change, so how it has played out is not yet known (Sello, 2025; Sello et al., 2024). My reading is that simplifying one requirement helps little while the bank account, the clearances and the premises remain.
Practitioners: qualifications without a career
The workforce is large, female, poorly paid and unevenly qualified. Adendorff and Bipath (2025) cite the national census as showing that only 52 per cent of practitioners hold a relevant qualification, with Level 4 as the usual entry point. Matlala and Molokwane (2024) cite the same census for the finding that close to half of practitioners have not had adequate training, and report stipends ranging from about R5,000 for Level 4 in the Free State to about R10,800 for Level 6 and above in the Western Cape. Those numbers come from secondary sources within the articles, so they are best read as indicative. Still, they show a sector where qualifications and pay are only loosely connected.
Rural and urban comparisons make the point sharper. Zulu et al. (2022) studied two Durban centres, one classed as rural in an informal settlement and one urban, with ten participants in all. The rural practitioners had little or no ECD qualification, worked from about 07:00 until around 18:30 with no paid sick or maternity leave, and mostly could not afford courses on their wages. Urban practitioners had more reasonable hours and living wages, and saw themselves as professionals. Both groups said parents treated them as babysitters (Zulu et al., 2022). A tiny sample, certainly, but it fits the national figures on pay and qualifications that other authors report (Matlala and Molokwane, 2024).
Status is part of the problem. Matlala and Molokwane (2024) cite the education department as putting women at 99 per cent of the sector, and describe providers as mostly African women from disadvantaged backgrounds. Adendorff and Bipath (2025) argue that the sector’s long reliance on low-skilled women, and the view of the work as simply nurturing, keeps its image low and puts off new entrants. If that is right, qualifications alone cannot lift the profession. Better pay and recognition would probably have to arrive alongside any new qualification.
The policy answer so far is a ladder of higher education qualifications, and the evidence suggests the ladder has few rungs at the bottom. Adendorff and Bipath (2025) surveyed 184 practitioners in Tshwane, Johannesburg and Ekurhuleni, and 95.5 per cent said they would take up the new qualifications. Funding was the barrier most often named, admission rules blocked many Level 4 and 5 holders, and recognition of prior learning was not widely applied. There was no sustainable employment or standard pay for the qualified. The authors argue that the policy will not work unless qualifications are linked to jobs and wages (Adendorff and Bipath, 2025). I think this is the strongest finding in the literature. Willingness to study is plentiful, and the missing piece is a funded career.
Quality and access in poor communities
Quality is where poverty shows up most clearly. Aina and Bipath (2022) compared four registered centres in suburban and township parts of Pretoria and found that practitioners in both understood why space, safety and play equipment matter, but township centres had fewer resources, shared toilets between staff and children, and lacked equipment for the youngest children. Suburban fees were close to four times the township fees, which paid for the difference. The authors call ECD’s promise as a great equaliser a smokescreen and ask how some township centres were registered at all, hinting at weak district supervision (Aina and Bipath, 2022). The Durban centre in the informal settlement had no learning materials, secure cupboards or proper toilets (Zulu et al., 2022).
Does quality matter more than getting children enrolled? Venter (2022) built a simulation of the Western Cape, a province the author describes as relatively rich and functional. The base case put about 60 per cent of five-year-olds in the poorest three quintiles as school-ready, against about 81 per cent in the richest two. Raising programme quality lifted the poorer group to roughly 70 per cent, full enrolment to about 65 per cent, and a large cut in community poverty to around 70 per cent. None of the three closed the gap (Venter, 2022). It is a model, with assumptions, and it should be read that way. Even so, it supports the view that more places will not do much unless the places are better, and that poverty itself sets a ceiling.
Access is shaped by the same forces. Moses and van der Berg (2023) used the 2019 General Household Survey and earlier ECD census data and found that household socio-economic status, living near a centre and the mother’s education were positively associated with enrolment, while having an unemployed or economically inactive adult in the household was negatively associated. They also found that five programmes gave surprisingly large learning gains for mostly poorer children, even where contact time was only two and a half to eight hours a week for three of them. Wills and Kika-Mistry (2022) add the fee story: about 39 per cent of adults living with children aged 0 to 6 said a child attended an early learning programme in February 2020, a share that fell to 7 per cent in mid-2020, and the recovery in 2021 was linked to families’ perceived ability to pay fees.
For children with disabilities the barriers stack up further. Karisa et al. (2022) cite figures showing that participation among 0 to 4-year-olds rose from 8 per cent in 2002 to 44 per cent in 2018, yet more than a million children aged 3 to 5 were still outside ECD centres. They argue for localised services, home-based and mobile programmes and better-trained staff, since poverty and transport costs keep many families away. The review contains no new data and little on funding. What it offers is a reminder that the function shift will be judged by whether the poorest and least visible children are reached (Karisa et al., 2022; Moses and van der Berg, 2023).
Conclusion
The move of ECD to Basic Education is a sensible administrative step, and the studies reviewed here do not suggest it should be reversed. What they do suggest is that its effects will depend on decisions that sit outside the shift itself. The subsidy is tied to registration, and registration to bank accounts, premises and clearances that poor centres struggle to meet (Sello, 2025; Sello et al., 2024). Practitioners are being asked to qualify without clear employment or pay (Adendorff and Bipath, 2025). The department is not the employer (Matlala and Molokwane, 2024).
Three priorities seem to follow. First, link funding to a costed plan that reaches unregistered and township centres, perhaps through conditional registration and a subsidy rate that tracks real costs. Second, attach qualifications to funded posts and minimum pay, so the ladder leads somewhere (Adendorff and Bipath, 2025; Zulu et al., 2022). Third, judge the shift by quality and access in poor communities and not by enrolment totals, since the simulation and census evidence points that way (Venter, 2022; Aina and Bipath, 2022).
The evidence base has gaps that should temper all of this. Several studies come from Gauteng, the best model covers only the Western Cape, and the census data describe a regulatory setting that changed in 2023 (Matlala and Molokwane, 2024; Venter, 2022; Sello, 2025). Rural centres appear mainly through one small Durban comparison (Zulu et al., 2022), and no study here follows children from a centre into Grade 1. Until the new department publishes costed plans and tracks outcomes, claims about what the shift has achieved will rest on hope more than on data.
Reference list
Adendorff, Z. and Bipath, K. (2025) Higher education qualifications for early childhood educators: policy implementation challenges. South African Journal of Childhood Education, 15(1), article a1597. doi:https://doi.org/10.4102/sajce.v15i1.1597
Aina, A.Y. and Bipath, K. (2022) Availability and use of infrastructural resources in promoting quality early childhood care and education in registered early childhood development centres. South African Journal of Childhood Education, 12(1), article a980. doi:https://doi.org/10.4102/sajce.v12i1.980
Karisa, A., Samuels, C., Watermeyer, B., McKenzie, J. and Vergunst, R. (2022) Priorities for access to early childhood development services for children with disabilities in South Africa. South African Journal of Childhood Education, 12(1), article a1119. doi:https://doi.org/10.4102/sajce.v12i1.1119
Matlala, L.S. and Molokwane, P. (2024) New home for early childhood development in the DBE: implications for ECD practitioners? South African Journal of Childhood Education, 14(1), article a1566. doi:https://doi.org/10.4102/sajce.v14i1.1566
Moses, E. and van der Berg, S. (2023) Evaluating the demand, supply and impact of early childhood development programmes in South Africa. Development Southern Africa, 40(6), pp. 1153-1173. doi:https://doi.org/10.1080/0376835x.2023.2198555
Sello, M.V. (2025) Determinants of non-profit registration in early childhood development centres in South Africa. South African Journal of Childhood Education, 15(1), article a1730. doi:https://doi.org/10.4102/sajce.v15i1.1730
Sello, M.V., De Wet-Billings, N., Mabetha, K. and Makuapane, L. (2024) Removing barriers to registration for early childhood development centres. South African Journal of Childhood Education, 14(1), article a1519. doi:https://doi.org/10.4102/sajce.v14i1.1519
Venter, L. (2022) A systems perspective on early childhood development education in South Africa. International Journal of Child Care and Education Policy, 16, article 7. doi:https://doi.org/10.1186/s40723-022-00100-5
Wills, G. and Kika-Mistry, J. (2022) Early childhood care and education access in South Africa during COVID-19: evidence from NIDS-CRAM. Development Southern Africa, 39(5), pp. 781-799. doi:https://doi.org/10.1080/0376835x.2022.2028607
Zulu, P.P., Aina, A.Y. and Bipath, K. (2022) Education and training experiences of early childhood care and education practitioners in rural and urban settings of Durban, South Africa. South African Journal of Childhood Education, 12(1), article a1167. doi:https://doi.org/10.4102/sajce.v12i1.1167