What is the best way to fund South African university education?

October 22, 2015


By Stuart Theobald[2]


Student protests across South African have made it clear that there is something very wrong with the current system of funding of university education in South Africa.

In an ideal world how would we want the system to work? How do we create something that ensures that every capable individual in South Africa is able to study and fulfil their potential, in a way that is fair to the aspirations of everyone else in the country?

This paper provides a brief consideration of the possible funding approaches and what might be optimal for South Africa. My purpose in writing it is to assist in thinking about possibilities that are feasible for the country.

There are three main options for funding university student education:

1. Funding out of the broad tax base (call this “tax funding”)

2. Funding up front by students, typically through fees while studying (call this “upfront funding”)

3. Deferred charges – students pay at some point in the future either through loan repayments or a tax on graduates

In reality no university system is completely devoid of tax funding, so the second and third options are really about funding sources in addition to tax funding. Within these three options, scholarships and bursaries may also have a role which I will discuss.

I will consider the issues regarding each of these in turn. But first some policy considerations so we can be clear about what outcomes are important.


Policy considerations

South Africa suffers extreme inequality, pockets of significant and persistent poverty, and a highly racially biased distribution of economic resources. Any policy needs to support efforts to eradicate this legacy, and higher education funding is one important mechanism that can be used. Young people who obtain university degrees see significant long run benefits to their life opportunities. University education funding is therefore a very important mechanism to ensure patterns of economic distribution don’t simply sustain from one generation to the next.

South Africa needs significantly more skilled individuals to meet broad economic development objectives. The National Development Plan says that we need to increase the number of students in higher education to 1.62m from approximately 750 000 now, and produce at least 5 000 doctoral graduates per year (from 1 420 in 2010) to achieve its vision 2030. Any funding strategy needs to ensure it supports the requisite growth in student numbers that is required to meet these goals. A funding mechanism that ensures capable students are able to access education and acquire such scarce skills therefore has significant public benefits. In addition to providing the necessary skills to drive the development of our country, universities are also tasked with conducting research which has significant public benefits too. Universities must address the research needs of the country ranging from economic and industrial development to health and welfare of all citizens. A funding mechanism therefore also needs to ensure that this important research function is supported.

Access to education is a right enshrined in our constitution. This applies to all levels of education including early childhood development. Primary and secondary school education is related to other policy considerations including widening access at university level, in that tertiary education success is substantially improved by high quality schooling at earlier levels. To the extent that there can be said to be a right to university education, this is restricted only to those who meet academic eligibility criteria. A weak primary and secondary education system is therefore an indirect mechanism to restrict access to universities because matriculants fail to meet admission criteria. Funding decisions therefore need to ensure that the appropriate allocations are made at each level to maximise outcomes.

There are many other important considerations that are relevant, but I consider those three – addressing inequality; providing the skills that SA needs; and supporting the right to access – the most important in any higher education funding debate.

Next I consider each of the three funding mechanisms available.

1. Tax funding

Universities currently receive a significant amount of funding from the national budget. In the 2015/16 national budget, R26.2bn was transferred to universities from national government to cover both operating and capital development costs. This will increase to R29bn in 2017/18 which will include R3.2bn to build two new universities, one in the Northern Cape and one in Mpumalanga at which point National Treasury foresees enrolment reaching 1-million. That implies funding of R29,000 per student, though this figure has been in decline as enrolments have been increasing. The total subsidy for universities is budgeted to grow by 5.7% per year in the medium term expenditure period. The universities allocation represents 42% of all of central government’s spending on post-secondary school education, which also includes all further education and training colleges that focus on vocational skill development. The figures exclude funding of the National Student Financial Aid scheme (NSFAS), which I will discuss further below.

The university subsidies have certain restrictions applied to them including specific allocations for capital expenditure and research. So the amount of university funding should not be seen entirely as a student cost subsidy, though some portion of it is.

In addition to the central government transfer, there are various funding provisions made by government departments to support research or specific skill development objectives. The state-owned organisations also fund various research and skills projects. However, for the purpose of general arguments regarding tax funding, it is central government transfers to universities that are the main focus of debate.

The central government transfer is an important source of funding for all public universities. For example, in 2014 Wits University received R1.09bn in state funding out of total revenue of R3.1bn (35%). It also received R1.28bn from student fees (41%) with the balance made up of sales of goods and services and other grants and income. Similar figures are found in the income statements of the other main universities. The trend has seen the subsidy declining in proportional terms over time. This has happened because universities have seen cost increases in excess of the growth in the subsidy, driven by increasing enrolment numbers and inflation. In order to meet those increasing costs universities have become increasingly dependent on student fees.

Full tax funding of higher education would see a substantial increase in the subsidy in order to eliminate the student fees line from the income statement.

There are positives and negatives in tax funding. I start with the negatives[3]:

1. Tax funding leads to a decline in student numbers or education quality, or both. Fully funded free university education could be achieved in an instant if university student enrolment was substantially reduced such that current subsidies fully covered student costs. Obviously that would be an undesirable outcome, but it illustrates that there is a tension between student enrolment numbers and the quantum of the tax subsidy. If student costs are fully funded out of the subsidy, there is no variable income to the university for increasing student numbers. Increasing numbers can only be done by a) an increase in the central government grant or b) a decrease in education quality by increasing class size, reducing administrative support, decreasing capital funding, or other cost reductions. Given that any central government allocation is constrained by competing demands facing the government budget as a whole, and the need for predictability in budgeting, there is an unavoidable tradeoff between student enrolment size and education quality within a tax-funded environment.

2. Tax funding harms access to universities by the poor. It may appear that the opposite should be the case, but the international evidence is clear that university enrolment is dominated by the children of the middle class. When the number of places is constrained, as is the case in any tax-funded environment, the middle class compete for access to those limited places, pouring resources into, for example, ensuring top high school grades. On any mechanism to allocate limited places, the middle class are able to crowd out the applications from the poor.

3. Tax funding is regressive. Graduates have much better life prospects than non-graduates and are much more likely to join the ranks of the middle class, even if they come from poor backgrounds. Any budget reallocation to university subsidies comes out of the national budget which must simultaneously support myriad other budget demands including increased allocations for welfare and poverty reduction programmes. While higher education funding has seen only modest growth in South Africa, growth in welfare and earlier education levels has been extensive, in part reflecting the government’s progressive priorities. Tax funding constrains other measures that may better serve the policy objectives of dealing with inequality and poverty in South Africa. It amounts to a transfer of wealth from the poor to the middle class. This point applies both to the opportunity cost of tax-funded university costs – i.e. that some other competing budget demand must suffer – but also the inequity that tax collection will be higher for non-university educated tax payers than otherwise. The taxes of an unskilled construction worker with no prospect of a university education would be used to fund young people who will be joining the middle class. Note that the regressive feature depends on the general background level of inequality and the proportion of graduates in a country. Very equal highly educated societies can fully fund university education without it being a disproportionate burden on low income earners and those without degrees.

One counter argument is that progressive basic tax structures eliminate this subsidy because the future wealthy pay higher percentage tax rates. However, this fact does not contradict the point that allocations to other poverty alleviation efforts are constrained by tax subsidies to universities. The fact that a private school matriculant with a bright future in investment banking will pay large amounts in tax one day is not a good argument for why his education now should be subsidised at the expense of other programmes.

4. Tax funding leads to inefficient student choice and supply inefficiencies. Students are better at determining what studies will further their ambitions than governments are. Student choice is therefore an efficient driver of university decisions over what courses to offer and invest in. While student fees are a limited source of income for universities, they do provide an incentive to drive universities to ensure that their offerings are attractive to students, as well as the non-university private education sector. When students are not paying fees they lose the economic power to affect supply-side decisions. This is true whether the students are meeting fees out of scholarships, bursaries or family money. The private higher education sector, including many colleges currently providing study support for Unisa degrees, would be effectively put out of business if the main universities become fee free.

There are three clear arguments in favour of tax funding:

1. Tax funding satisfies any claims to have a right to higher education. Such a right is not, on its own, a good argument that the state should pay for it. Everyone has a right to freedom of movement, but that does not mean the state should be liable for everyone’s taxi and air fares. The state has a responsibility to ensure access, but this does not mean it must itself pay for it. If the state can ensure the right is met, without having to pay the bill itself, then that is obviously preferable, given that state resources are finite.

2. Tax-funded models are simple to administer and understand. In models such as that applied in Germany, the universities do not have to carry the cost of administering student fee collection systems. It is also much easier for young people to connect their goals to their choices, for instance knowing that the only barrier to university access is getting good enough grades, instead of having to worry about financial restrictions. However, because in South Africa differential primary and secondary schooling is a massive determinant of access to, and success at, university, access may be better promoted by focusing on funding better primary and secondary schooling than on universities.

3. There are significant public benefits that graduates provide including the application of their skills in the economy at large. But there are also significant private benefits to those graduates, such as their increased earnings capabilities and better quality of life that degree educated individuals may experience. The fact that a public benefit exists is not always a reason the state should be required to pay for it – many public benefits are by-products of private benefits, for example the availability of free emergency calls from cell phones. If the public benefit exists as a by-product of individuals’ pursuit of private benefit, and that private benefit is sufficient to motivate the choice, then the public should take the free ride.

Given the policy objectives outlined at the start, these arguments provide strong theoretical reasons we should not support 100% tax funding. Tax funding does not support the objectives of eliminating poverty and inequality and may in fact worsen it. The need to expand the numbers of university enrolments while protecting quality advises against it. And government’s obligation to meet the right of access to education does not imply it should fund it, if it can find alternative ways to ensure access happens.

It is clear then that tax-funding should be supplemented by either upfront fee payments or deferred charges, provided these are not a barrier to access for deserving students. In what remains I consider these alternatives.

2. Upfront payments

These typically refer to fees charged by universities, a proportion of which is often due at the start of any programme. Fees can be at various levels and make up various proportions of universities’ income. As discussed above, for example, Wits University receives 41% of its income from fees. Basic annual fees range from R26,016 at the University of the Free State to R46 000 at the University of Cape Town. Increases in these fees have been a catalyst for the nation-wide protests we are currently experiencing.

The arguments over upfront payments are clear. If students cannot afford them they are simply prevented from access. This means universities will then be dominated by students from well-resourced backgrounds whose parents are able to pay the upfront fees, or older students who have been able to generate sufficient savings to meet those fees. Some students may be able to access private funding, such as student loans from the commercial banking systems, but such loans usually include guarantees from parents and some portion of own resources to limit the risk to such institutions. Commercial interest rates are comparatively high. Alternatively, some students may be able to get scholarships and bursaries, depending on whether their study objectives match those of companies or other funders.

Requiring students to work up front to generate savings from which to meet fees is inefficient and, given the massive youth unemployment problem in South Africa, unrealistic. The inefficiency considerations are similar to buying a house: people could either work for a long period of time to save up to buy the house, or take out a 20-year loan to buy it now. Provided interest rates are not punitive, long-term welfare is maximised through the loan. This argument is even stronger in the case of university education because the earning power of the graduate is higher after the degree than before it.

Commercial funding, even in a market with a relatively efficient banking system like South Africa’s, is not a sufficient answer. To be eligible for such funding students and their families still need assets and income that will satisfy the collateral needs of the funder. This means commercial funding is only suitable for prospective students from the middle classes.

The one positive argument in support of fees is that there are efficiency gains. Students are best positioned to determine which financing option will have the least impact on their lives. Some students may have high earning power for whom saving before education is preferable. Others will want to pursue studies that company sponsors are willing to fund. Others will have resources that provide them with good access to commercial finance. Still others will have wealthy parents. These sources of funding, for those particular students, may be much more efficient from an overall point of view than using government funding, given the attendant opportunity costs on the public at large. If students have options other than government funding, they should use them.

A related argument is that fees maximise choice efficiency. Students choose degrees that will enhance earnings such that the upfront fee and costs of finance represent a good investment. However, this argument is weak. The long-term returns to degrees are often difficult to predict so students will choose only low-risk options such as professions like accounting, medicine, and so on. Given that the future is uncertain, we should want graduates to take risks that provide flexibility to adapt to a changing environment. Forcing them to, in effect, self-insure, discourages certain desirable behaviours such as the willingness to experiment. This argument can be taken too far though – we can conceive of degrees that have no value whatsoever of either a private or public nature, and we must be sure that whatever optimal system is designed these are discouraged.

Given the policy considerations outlined at the start of the paper, upfront charges are clearly not optimal. They prevent access to those without financial means, entrenching inequality. They also prevent expansion of enrolment given that fewer prospective students will be able to afford access. But, where upfront funding is more efficient for the reasons outlined above, it may be preferable. This will certainly only apply in the case of certain students who have existing financial means or access to commercial funding because a company perceives private benefits to itself. A funding solution therefore should not displace these more efficient alternatives.

Given that upfront fees are substantially exclusionary, they are clearly not suitable for most students. For those students, deferred charges, in the absence of full tax funding, are a better option.

3. Deferred charges

Deferred charges refer to mechanisms where the costs of university education are met by the student at some point in the future. Such systems are used in various countries and are usually income contingent, so students pay a portion of their income over some thresholds toward recovering some part of the costs of their education. They can be structured either as loans that are repaid by students or as a graduate tax.

South Africa has a variety of this approach in the National Student Financial Aid Scheme. In the last budget allocation, it received funds of R9.2bn. According to its last annual report in 2013 it provided some funding to 416,174 students, including 194,923 at universities. Funding is provided at the repurchase rate less two percentage points, which is significantly below commercial rates as well as the government’s own cost of borrowing, representing a hidden subsidy by tax payers. Repayments take place once students have begun work and are earning at least R30 000/year, with repayment rates increasing with salary tiers. Applicants are means tested and only those from the poorest background are eligible for funding. Children from middle class backgrounds, for example those of civil servants, are unable to access this funding and often fall into a funding gap.

The NSFAS system is beset by problems. In 2014 it distributed funds of R8.6bn, but collected repayments of only R300m. That alarmingly low collection rate reflects the growth in disbursements, inefficiencies, and the design features of the scheme, but it also clearly indicates that it has drawn massively on the budget resources. Therefore the NSFAS is more like a subsidy scheme than a cost deferral scheme in practice. The central budget allocated R9.2bn to the NSFAS in 2014/15.

A 2010 ministerial review of NSFAS identified major problems in the way it works. Of the students who it funds who have left university, 72% drop out without a degree, making repayments difficult. The reasons for that poor performance are many, including that the funding is insufficient to allow students to fully focus on their studies. The weak family support structure for such students, given that they often come from environments with little experience of tertiary education, is part of the problem, as is the quality of their preparation in high school. Students are also financially vulnerable in that NSFAS funding seldom covers the full cost of tuition or living costs. NSFAS was also beset by administrative problems, receiving qualified audits and battling to administer applications and track borrowers.

Another feature of the NSFAS system is that loans are converted to bursaries in the event that the student graduates. So that means that graduates with the greatest ability to repay their funding do not have to, while those who drop out do. Given the policy objective of reducing inequality, this is regressive.

Some steps have been taken to address the myriad problems with the NSFAS, though many remain. Among the proposals in the 2010 review was that the SA Revenue Service take over the collection of repayments so that it functions in effect like a tax. This proposal is apparently still under discussion.

Whatever the problems with the NSFAS in particular, the general features of deferred cost models are:

1. University is free at the point of use. Deferred fees means there is no payments by students at the time they are attending university. This means there is no barrier to access posed by fees.

2. Deferred charges can be income contingent. Repayments depend on how much the graduate is actually earning. This is very different to a normal student loan which would require payments of X amount over X number of years. Income contingent repayments can be scheduled to match an earnings cycle. There is therefore a built in mechanism to cope with those graduates who cannot in future repay. This in effect amounts to an insurance scheme that transfers the worst aspects of the risk of future repayments away from the graduate.

3. Deferred charges can extend to cover living costs. When structured as income-contingent loans rather than a blanket university-funding graduate tax, deferred charge models can also be used to finance students’ reasonable living expenses during the period of study. Given that these will be repaid at some point in future, the incentives remain for students to not live exorbitantly and to use own resources if available, so being efficient with government funds.

Such deferred cost funding models based on income-contingent repayments are progressive in that they are responsive to different future earnings levels. They also ensure that there is no financial barrier to students while they are studying. They meet the policy objectives of addressing inequality, widening access to university, enhancing skills levels, while also ensuring the opportunity cost of university funding is not born by the poor and tax payers who do not have degrees.

There are two approaches that can be taken: loan repayments or a graduate tax. Loan repayments are finite contributions to meet some specified amount based on the actual cost of their studies.  A graduate tax is an amount paid by all graduates in addition to their income tax, though some caps on the amounts can be applied. As typically applied, graduates form a different class of taxpayer which in effect carries an additional tax burden. Both are income contingent. The issues relevant to these alternatives are:

1. Graduate tax is blind to the actual cost of the choices students make on what to study and how to fund it. It is therefore inefficient: students are incentivised to choose the most expensive course of study (assuming correlates with benefits) and to claim the most possible for living expenses.

2. Graduate tax is insensitive to whether students have chosen to incur upfront charges, perhaps because they have scholarships or bursaries from companies, have own resources from their own work, or their parents have funded them. It therefore discourages these alternative forms of education funding and increases the burden on every other tax payer.

3. Graduate tax models discourage private provider participation in the higher education sector. Given that graduate taxes would be used to fund universities centrally, only public universities would receive the benefit while private providers would have their business models undermined. Private providers can be efficient providers of education in some areas, particularly vocational training. Private providers should not receive government funding directly because student choice is critical to driving their efficiencies.

4. Graduate tax can only be collected from those who stay in the tax net. Graduates who immigrate will not carry any burden for the benefits they receive. Those who exit the tax net, for example if a spouse covers their living expenses or they are beneficiaries of an inheritance or other non-taxable receipt, will also escape any obligation.

5. It is harder to justify living costs as part of a graduate tax scheme. Students need to have their living costs covered while they are studying. Students can naturally be economical or not in their living arrangements and delinking these decisions to the amount they repay is difficult to justify.

6. A graduate tax is incurred no matter what the public benefit features of students’ degree choices. Such taxes are incurred by all graduates as a function of their salaries, even if they made choices to obtain skills for which there is large public benefit.

The alternative of income contingent loan repayments exhibit these features:

1. Loans can be converted to bursaries where the public benefit gains justify it. A loan scheme can build in conversions in the case of education choices that exhibit clear public benefits, such as in the case of scarce public sector skills. This is a decision that can be made upfront to guide student decisions. In contrast, providing graduate tax exemptions is post hoc and vulnerable to criticism of arbitrariness.

2. Loan payment amounts are a function of the cost of education. This creates incentives for students to be economical with their education choices.

3. Loans can cover tuition costs and also living costs. Because there is a direct link between loan obligation and living expenses, it is more clearly feasible to justify the extension of loans to cover living expenses.

4. Loan recovery can be shaped such that graduate emigrants still carry the liability. Those who take their skills to other countries will still carry the burden of the costs of their education.

5. Income contingency means there is no excessive burden on borrowers. The repayment amounts can be structured to only apply above certain income levels. Cumulative capitalised interest can be capped at a point such that the state effectively insures borrowers against the repayment consequences of low incomes.

To some extent a graduate tax can be shaped to act like a loan scheme and vice versa. The considerations above apply as such schemes are typically implemented and it is to a large extent the arguments, rather than the particular description of one as a tax and the other as a loan, that matter.



These considerations lead me to conclude that the following policy approach is optimal:

1. A tax-funded subsidy is a critical component of the overall cost of student education. The quantum of the subsidy is a function of political trade-offs. One important principle is that the subsidy should be such that the marginal cost to the university of accepting one more student is the same as the variable income from that student. In other words, universities should not incur financial harm in expanding enrolment. This means subsidies should focus on research and capital costs and less on the variable cost of student enrolment. There should still be a material subsidy for student costs, reflecting in part the public benefit from university educated citizens.

2. Those who can pay upfront fees, should. Whether out of family resources, income, company-provided bursaries, or funder-provided scholarships, students who have access to these means should use them. This lessens the burden on government funds and is progressive in nature, ensuring government funding support goes to lower-income beneficiaries. It is also efficient in that students access the lowest cost sources of funding.

3. There should be zero upfront costs for students who cannot afford them. A deferred cost model is preferable. This should include covering of students’ living expenses where they cannot afford them. Repayment should be income contingent, such that students pay a percentage of their income above certain thresholds. The scheme should be structured such that the obligation remains for those who exit the tax net, but falls away for long-term low-income earners through a form of implied insurance.

4. In cases of particular public benefits, loans should be converted to bursaries. Many study choices are made because students desire the private benefits that accrue to them. In some cases the public benefits are particularly substantial and the state should encourage these with targeted bursaries. That is feasible in both an upfront and deferred cost model, but impossible in a tax-funded model. In a tax-funded model, government can direct resources to particular subjects with public benefits, but it is preferable to leave universities to offer choices incentivised by the marginal revenue from each student, and to equip the students with such marginal revenue through scholarships and bursaries, and deferred cost schemes.

It follows that government funding should remain an important part of the overall funding equation. The declining proportion of government subsidies in overall university funding is regrettable. Such funding should at least keep pace with the cost increases universities face. It is also desirable to increase funding proportional to student enrolment figures, though if deferred cost models meet the marginal increase in costs from each additional student, this may not be necessary.

Most importantly, government needs to facilitate an efficient and effective deferred cost model that meets the objectives outlined above. The National Student Financial Aid Scheme is probably the institution that should be built into the ideal model, with support from other arms of government. The SA Revenue Service should administer loan recovery. Bursaries should only be provided for specific scarce skills with a clear public benefit. Means testing should be generous enough to include all borrowers who cannot access commercial funding. Repayment should be income contingent with a write-off period for those who do not find employment or earn salaries below a threshold.

Of course the NSFAS is very far from such an ideal institution, but the political will generated from current student protests will hopefully contribute to changing that. Encouragingly, the private sector has recently taken steps to support the institutional capacity of the NSFAS and there are various government initiatives to support it.

Politics is not determined by rational considerations of optimal public policy. Students’ short-term needs can clash with the governing party’s other political objectives. Also, government capacity can fail us, even when the optimal policy route is clear. Those are problems we must all find our way through. I hope this paper provides a walking stick along the route.

[1] I call this a working paper as a way of saying it is not at the level of passing full academic muster. It has been written to contribute to the debates about higher education currently gripping South Africa. I’ve prioritised getting the arguments together ahead of proper referencing. I provide some references in the text and footnotes. It also will evolve as I receive responses and have more time to compose supporting data. Perhaps it will become a final paper in time.

[2] I am chairman of research house Intellidex. I am also a part-time PhD student at the London School of Economics, though working on the relatively unrelated field of philosophy and theoretical finance. This paper is written in neither of those capacities. My undergraduate education was at Rhodes University.

[3] Most of the economic arguments in this paper have been inspired, if not taken wholesale, from the work of Nicholas Barr. I have not provided full referencing but his papers Financing Higher Education, in Finance and Development, June 2005 , Volume 42, Number 2; and Financing higher education: Comparing the options published at
Other references are in the text.