Fact and Fiction - A comment on the 2025-6 Budget
- Professor Imraan Valodia
SA's overall economic strategy in the budget and the proposed VAT increase demand more rigorous debate.
The 2025-26 annual Budget has been subjected to an unprecedented degree of interrogation and deliberation. But two inter-related issues have not been debated with sufficient rigour: the overall economic strategy in the budget, and the proposed increase in Value Added Tax (VAT).
The VAT debate has been based largely on conjecture and ideology, rather than on the actual evidence of the impact of VAT, and its role in the country’s larger economic strategy.
I address these lacunae in the budget debate in two parts: Part A addresses economic strategy. Part B is focused on evidence on the impact of VAT in South Africa,
Part A: The Budget and economic strategy
There is widespread agreement that the key challenge facing South Africa is how to shift the economy to a trajectory of higher growth. Some would argue that any growth would be good, because ultimately the benefits would trickle down to low-income groups. The evidence does not support this view: there is now acceptance across the world, including in South Africa, that ‘trickle down growth’ is a mirage. Moreover, it is politically unstable. South Africa needs a growth trajectory where the benefits of growth disproportionately benefit low-income groups, rather than causing further unequal concentrations of wealth.
South Africa also has a debt problem. In 2025-6, 22% of total budget expenditure will be spent on servicing debt – simply paying the interest on what we have borrowed in the past. To put that into perspective, we will spend more on debt servicing than on basic education, the next largest slice of government expenditure.
To try to deal with the debt issue, government has over the last few budget cycles followed a strategy of austerity, by reducing the rate of growth in social expenditure, and forcing a reduction in real per-capita allocations, especially to vulnerable populations.
It is a good idea to cut wasteful expenditure, which has increased substantially over the past decade. In the past 10 years, South Africa has spent some R520 billion to bail out state-owned enterprises, many of which are either now defunct or in no better state.[2]
A key element of expenditure that has made the current budget more complex is government’s cavalier approach to the public sector wage bill. Just three days before the aborted budget speech, government announced a 5.5% increase in the wages of public servants – significantly above the rate of inflation. In terms of payments to third parties, the largest item of expenditure in the budget is compensation to state employees: for 2023-4, we spent some R723 billion on the government wage bill.
While it is true that public servants should be well paid, it is also true that we have a problem of productivity in the public service. One key question for the budget is this: how is it possible that given our challenges, government agreed to increase its largest item of expenditure at above the rate of inflation? This seems inexcusable. Moreover, the public sector wage agreement has implications for other parts of the labour market – in the private sector and in public institutions such as universities – because it serves as a benchmark for other wage settlements.
Because government has not been able to reprioritise key expenditure, it has instead followed a programme of across-the-board cuts, especially on health, education and related social services. As my colleagues at SCIS have repeatedly pointed out, this austerity programme has significantly undermined the social fabric of South Africa, hurt the most vulnerable groups in our society and, paradoxically, not resulted in a credible debt consolidation programme.[3]
In the 2025-6 budget commentary, many have called for further radical cuts in expenditure to reduce the gap between tax revenues and expenditure, without increasing taxes. Although there is wastage in government, a focus on waste alone is not going to solve the debt or the growth problem. There are several reasons for this. First, expenditure reviews and action to address waste take some time to have effect and, whatever the politicians may say, these will be met with political resistance. In short, it’s not going to solve the problem in the immediate term.
Second, we should be careful that quick and radical cuts to expenditure do not further undermine our economic and social infrastructure, which is badly in need of maintenance, renewal, and expansion.
Third, cutting expenditure in targeted areas, such as state-owned enterprises and related institutions requires political will. It is important to note, as Andrew Donaldson has demonstrated,[4] that even in the 2025-6 Budget the largest increases in expenditure are in ‘extra-budgetary’ institutions such as PRASA and the Road Accident Fund.
Fourth, given that a large proportion of expenditure in South Africa is on social services – education, health, social support – across-the-board cuts will further burden the most vulnerable members of society and be regressive. As colleagues and I have argued elsewhere,[5] government expenditure and taxation have a huge impact on improving inequality outcomes in South Africa because the benefits of public sector expenditure accrue disproportionately to low-income groups, and taxes are largely paid by higher-income groups. Moreover, radical cuts in expenditure lead to poor investment decisions, and irrational outcomes. For example, South Africa correctly spends a large amount of public funds on training medical doctors. Yet even though we have a shortage of doctors in rural health services, many young medical graduates are unable to find work due to recent cuts in provincial health budgets.[6]
As a general principle, the government budget should be counter cyclical – when economic growth is low, public expenditure should increase to stimulate economic growth. This has not occurred over the past few years. Instead, we have had a slowdown in expenditure, which has locked the economy into a low growth cycle.
There is another option, which appears to be the plan in the original 2025-6 Budget proposed by the National Treasury. This had the objective of setting the economy onto a path that creates the possibilities for us to grow our way out of the problems that bedevil our economy. Most of the challenges on the debt side are expressed as a percentage of GDP. So, if GDP increases, many of the stress points on the debt issue are eased. Moreover, as GDP increases, tax revenues increase, so easing the deficit. It becomes a virtuous cycle. Thus, the trick is to find a way for the economy to grow in a sustainable way, and that was the plan that was evident in the Budget.
In summary, for some years now, government has been trying to address our debt challenge by reducing expenditure. This has not worked, and it has had detrimental effects on vulnerable South Africans. A better strategy is to use the budget to push the economy onto a sustainable growth path by increasing expenditure on social and economic infrastructure and social protection for the poor. This requires the Budget to find additional sources of revenue to finance an increase in expenditure. If this is achieved, the economy can grow, our social challenges can be addressed, and our debt is reduced. All of this requires some hard choices on the revenue side.
Part B: Revenue and the proposed increase in the VAT rate
A major controversy in the Budget has been the proposed two percentage point increase in VAT. Most commentators, including political parties opposed to the proposal and those claiming to represent low-income groups, have argued that an increase in VAT places undue burden on low-income groups, and is therefore regressive. Several alternatives have been suggested, such as higher corporate tax rates, a wealth tax, and cuts in expenditure.
There are three important issues about how taxes work. There can sometimes be a large difference between who pays a tax and who ultimately bears the real cost, or in economic terms, the incidence of that tax. Take the example of corporate taxes. One might levy a higher tax on companies, but if companies operate in markets that are not competitive, they may be able to pass on the full cost of the tax to consumers through higher prices. So, while an increase in company taxes might appear to tax profits in companies, at least some of this might be passed on to consumers, including those in low-income groups. This is not to argue against corporate taxes, but rather to demonstrate that we should comprehensively consider how taxes impact on society.
Second, an increase in tax, whatever the form, is associated with a behavioural change on the part of economic agents who may be affected by that increase. Let me demonstrate this with the issue of wealth taxes. My colleagues at the Southern Centre for Inequality Studies have developed a succinct and viable proposal for such a wealth tax, which I support.[7] However, the mere mention of a wealth tax leads wealthy South Africans and their tax consultants to shift wealth from South Africa to jurisdictions where such wealth can’t be subjected to such a tax. Similarly, increasing income-tax rates on the highest income earners results in changes to their remuneration packages. The key point is that generally, a new tax or an increase in the rate of an existing tax usually leads to less tax revenues than may be planned because of these behavioural responses. For several years government has not adjusted personal income tax for inflation – effectively increasing the rate. There is a limit to how much we can continue to do this.
Tax experts usually refer the three Es in taxes – equity, efficiency, and ease of administration – for evaluating tax policy proposals. In other words, any new taxes should ideally promote equity (they should be progressive, not regressive), they should be efficient, and easy to administer.
Many have argued that the increase in VAT places an undue burden on low-income groups and is therefore regressive. But this view is not based on the actual evidence of the impact of VAT in South Africa.
It is important to understand the meaning of the words “progressive” and “regressive”, in how we evaluate tax proposals. Let’s take the example of two families – Family A and Family B, respectively, a high-income and low-income family. Family A has a total annual income of R3 million and consumes 80% of its income and saves the balance. Family B, lives close to the breadline, has an income of R60 000 and consumes all of it. VAT is a consumption tax, so it affects only the income that a household consumes.
Assuming VAT is charged on all goods, and the rate is 15%, Family A will pay VAT of R313 043 (15/115 multiplied by the R2.4m Household A consumes), and Family B will pay R7 826 (15/115 multiplied by the R60 000 Household B consumes). So, government collects 40 times more VAT from Family A than it does from Family B. However, as a proportion of its income, Family A has an effective tax rate of 10%, while Family B has a tax rate of 13%. In this case, VAT is regressive, since as a proportion of its income, the VAT paid by Family B is greater that of Family A. But the assessment does not tell us how much tax different groups will pay. As the example here shows, though VAT is regressive, the rich household actually pays substantially more in VAT. It’s worth pointing out that Household B is below the threshold for personal income taxes (PIT), whereas Household A will also be subject to PIT.
According to the International Monetary Fund (IMF), VAT is now the mainstay of tax systems in more than 160 countries, raising on average one-third of total government revenues.[8] We should be concerned about the impact of VAT for at least two reasons. First, as illustrated above, VAT places a heavy burden on low-income households that spend a large proportion of their incomes, Second, VAT may also place a heavy tax burden on women. In South Africa, as in many other countries, female-led households tend to be clustered at the lower ends of the income distribution.
Furthermore, women usually take more responsibility for feeding children and caring for family members. Thus, they are more likely to spend a large proportion of their incomes on consumption goods such as food. So, in general, VAT is gender regressive, and, at least in theory, is a regressive tax. But is it so in practice?
Many commentators on the tax proposals, especially on VAT, have not bothered to look at the actual evidence. At least three studies have explored this issue in some detail. In 2008, I worked with colleagues in eight countries (South Africa, Ghana, Uganda, Morocco, Mexico, Argentina, India, and the United Kingdom) on the gender issues related to tax. A key concern was whether VAT is in fact regressive and placed an undue burden on low-income, and especially female-headed, households.[9]
Our findings are that, in general across the countries studied, VAT is regressive, but it depends on how it is implemented. In South Africa, because the zero-rating of basic consumption goods is very effective and protects low-income and female-headed households from the deleterious effects of VAT, it is not in fact regressive. Compared with other countries in our study, South Africa stood out as an example of how it is possible to have a VAT system that is neutral – that is, it is neither regressive nor progressive. A more recent study by South African economist Ingrid Woolard and colleagues[10] on the Davis Tax Committee [11] reached a similar conclusion in 2018: in South Africa, because of how effective zero-rating is, VAT is NOT regressive, but proportional. A third important study in this regard was done in 2018, when VAT was increased from 14% to 15%.
Following a similar emotive debate, the Minister of Finance appointed an independent committee, on which I served, and which was chaired by Ingrid Woolard, to advise on further zero-rating.[12] Our conclusion was that zero-rating is highly effective at protecting low-income groups from the deleterious effects of VAT. This held true for all except a few items, where we recommended extensions to the basket of goods that are zero-rated.
The challenge with zero-rating is that while low-income households benefit from zero-rating, it is not possible to exclude high income households from consuming such goods. In absolute terms, because they spend more, high-income households will benefit more from such zero-rating. The policy challenge is that, in trying to protect low-income groups, large amounts of potential VAT revenue are lost to high-income groups, who don’t need to be protected.
The trick is to find a basket of goods that low-income households consume in large quantities and that high-income households don’t. For these goods, the so-called gain ratio from zero-rating will be high. Some typical examples of such goods are samp, beans, canned pilchards, and cabbage. These are all goods that low-income households consume that high-income households do not. So, zero-rating them effectively protects low-income households, but there is no serious loss of VAT revenue from high-income households. A good example of the trade-offs with zero-rating is the case of chicken. While it is true that chicken is an important source of protein for low-income households, this is also the case for high-income households. So, if all chicken were zero-rated, this would protect poor households, but a large amount of VAT revenue would be lost because, in absolute terms, high-income households spend more on chicken than do low-income households. In our 2018 zero-rating report, at 2018 prices and consumption patterns, the three lowest deciles of income groups paid R1.3 billion of VAT on their purchases of chicken. In contrast, the other deciles spent R4.6 billion in VAT on chicken purchases. In other words, to provide relief to the low-income households equivalent to R1.3 billion, government loses R4.6 billion to high income households - not a good trade-off. However, some chicken products, such as chicken heads and feet, are disproportionately consumed by low-income groups, and are therefore good candidates for zero-rating. National Treasury’s proposals for increasing the basket of goods to be zero-rated are based on solid research and should be supported.
In addition to equity, the two other Es - efficiency and ease of administration - of taxes are key to consider. On these, VAT has big advantages. When evaluating the efficiency of a tax, our concern is how much of the potential tax is actually collected. As outlined above, we should consider the behavioural response to VAT. Many taxpayers find ways to get around their tax obligations, either through tax avoidance or tax evasion. The former refers to behavioural responses that are legal – such as high-income earners restructuring their employment packages to be more tax efficient. In contrast, tax evasion is illegal and outside the law. But it happens – and the less effective a tax authority is at clamping down on evasion, the more likely it is to occur. Because of the way that VAT works, it’s very difficult to avoid or evade (because it is collected along the chain of production, evading it or avoiding it along the production chain can be costly). One way to measure the efficiency of VAT is to assess the proportion that should be collected against what is actually collected. We can assess this by looking at the so-called c-efficiency ratio of VAT. Simply put, we can use the consumption aggregates in the GDP to calculate this. The Davis Tax Committee assessed this and concluded that South Africa does quite well on this score. Compared with our peers, the c-efficiency ratio is South Africa is high – this means that we have very little leakage in the VAT system. Finally, since South Africa has an effective VAT collection system, it is relatively easy to increase the VAT rate without needing to invest additional resources to collect the VAT – so the ease of administration is another positive for VAT.
Three issues are worth emphasizing on the VAT issue. First, and most importantly, contrary to what many commentators have argued, VAT in South Africa is not regressive. It will add to the burden of low-income households, but most of the VAT will be collected from higher-income households and zero-rating will go a long way to protecting the lowest-income households. Second, VAT is a very efficient tax. For relatively low increases in the rate, government is able to raise a large amount of revenue. Finally, the system is easy to administer and adds very little additional costs to ensure collection of the taxes resulting from the increase in the rate.
Conclusion
To address our problems of unemployment, poverty, and inequality, we must find ways for our economy to grow. The budget is among the most important instruments that government has to shift the economy onto a sustainable growth path.
A key barrier to growth is the difficult fiscal position we find ourselves in. The servicing of debt now consumes over one-fifth of total expenditure and is crowding out government’s ability to invest in economic and social infrastructure critical to shifting the economy onto a growth path. In recent years government has reduced growth in some areas of expenditure, mainly social expenditure, to deal with this debt issue. However, this austerity programme, while harming low-income earners who rely on government expenditure, has not provided a credible path to address the debt issue. More of the same is not going to solve our problems.
A more sensible pathway is to raise government revenue. This will allow government to make investments in social and economic infrastructure, stem the negative impacts of austerity, and shift the growth trajectory of the economy. This will also provide a credible path to addressing the debt issue. Hard but sensible choices need to be made. An increase in VAT is a good pathway to follow – it raises the revenue we need, allows government to increase expenditure in critical areas, and, as the evidence shows, it does NOT place an undue burden on low-income groups. The primary beneficiaries of the additional revenues will be low-income groups, so the total effect will be progressive. Other options such a reducing expenditure on health, education and social security will disproportionately hurt the poor.
Another key consideration for tax policy is that it must be politically credible. The population should believe that their tax contributions are being used effectively. If this is not the case, then tax morality becomes an issue – people don’t believe in government’s ability to spend wisely, resistance to taxes increases, tax avoidance and evasion increases – it’s a dangerous cycle.
It would be fair to say that, with the high levels of corruption in our political system, government’s credibility is low. Thus, if VAT is to be increased, and, as I have argued above, it’s the best option we have, government must improve its credibility and assure South Africans that tax revenues will be well spent. The population needs to be reassured that government will address rampant corruption, be more circumspect about wage policies, address the bloated and expensive political system, improve the level of and efficiency of government and, most important, address the high levels of waste in public expenditure.
Notwithstanding its critics, the proposed increase in VAT and the associated shift away from austerity in the 2025-6 Budget is good economics and is supported by research evidence. This should be supported by the citizens of South Africa. Government now needs to show that it is serious about good economic policies and effective use of tax revenues.
________________________________________
[1] Imraan Valodia is the Pro Vice-Chancellor: Climate, Sustainability and Inequality and Director: Southern Centre for Inequality Studies, University of the Witwatersrand
[2] https://businesstech.co.za/news/finance/814839/south-africans-kiss-r520-billion-goodbye/
[3] See, for example: https://wiredspace.wits.ac.za/items/904164f5-6e03-4b31-a1e3-10720e86293b
[4] https://saldru.uct.ac.za/articles/2025-03-31-political-economy-2025-budget
[5] https://www.witspress.co.za/page/detail/Inequality-Studies-from-the-Global-South/?k=9781776146161
[6] See, for example: https://bhekisisa.org/health-news-south-africa/2025-03-12-update-budget-2025-jobs-for-only-800-of-1-800-unemployed-doctors/
[7] /media/wits-university/faculties-and-schools/commerce-law-and-management/research-entities/scis/documents/WorldInequalityLab_WP2021_02_SouthAfrica_WealthTax_2b.pdf
[8] https://www.imf.org/en/Publications/fandd/issues/2022/03/b2b-value-added-tax-continues-to-expand#:~:text=It%20has%20become%20a%20major,of%20their%20total%20tax%20take.
[9] For the results of these studies see here: https://idrc-crdi.ca/en/books/taxation-and-gender-equity-comparative-analysis-direct-and-indirect-taxes-developing-and.
[10] https://www.commitmentoequity.org/wp-content/uploads/2017/09/CEQ-WP29_Inchauste-Lustig-Maboshe-Purfield-Woolard_South-Africa_REVISED-May2017.pdf
[11]https://www.taxcom.org.za/docs/20180329%20Final%20DTC%20VAT%20Report%20to%20the%20Minister.pdf
[12]https://www.treasury.gov.za/comm_media/press/2018/2018081001%20vat%20panel%20final%20report.pdf