The Political Economy of the 2025 Budget

South Africa’s finance minister faces opposition to his budget proposals, both in cabinet and in parliament. By longstanding tradition, spending and revenue proposals are presented on Budget Day as the consolidated fiscal plan of the government of the day, on the understanding that they will be supported through the legislature’s committee processes and passed into law later in the year. This presumption has been upended. Heightened parliamentary scrutiny of the public finances lies in prospect.

The national budget is unavoidably detailed and complex, and is the outcome of established institutional processes, political priorities, diverse interests and influences and extensive analysis and arithmetic. The integrity and coherence of the budget process depend on both the comprehensiveness of its coverage and a workable balance between “top-down” fiscal planning and prioritisation, and “bottom-up” compilation of expenditure and revenue estimates on huge numbers of programmes, activities, public entities, funds and departmental components. Despite the appeal of “zero-based” approaches to expenditure reviews, budget preparation is always, in the main, a structured consolidation of incremental adjustments to past plans.

Since 1997, continuity and transparency in South Africa’s budget and fiscal policy have been supported by publication of a Medium-Term Budget Policy Statement (MTBPS) several months ahead of the main budget in February each year. A three-year forward planning approach to both economic projections and budget estimates was also introduced at this time. Expenditure estimates have always been prepared taking inflation projections into account; in recent years the Treasury has also tabled its revenue projections on the assumption that excises and income brackets will in future years be adjusted for inflation.

Neither the 2024 Budget forward estimates nor the October MTBPS signalled that a substantial tax increase would be required in 2025. We now know that in the weeks leading up to the February budget the minister indicated to cabinet and the GNU partners that a VAT increase should be anticipated. The DA and several other parties indicated that they would not support this. When the tax proposals were shared with cabinet on the morning before the budget, several ANC ministers also voiced reservations about the proposed 2 percentage point VAT increase.

The tax proposal has to be viewed in the context of the broader economic outlook and government’s spending plans.

The macroeconomic position is dire. In the 2024 Budget, growth for the year was projected at 1.3%, by February the Treasury’s estimate was 0.8% and the outcome reported in the latest GDP statistics was 0.6%. GDP inflation has also come in below expectations, and overall nominal growth in GDP last year was just 4.4%, down from 7.1% and 5.5% in 2022 and 2023 respectively. Gross fixed capital formation (investment) growth between 2023 and 2024 was 0%, which places some doubt over the treasury’s projected 1.9% growth for 2025.

If the problem was cyclical, these would be circumstances that would call for an expansionary fiscal stance. But there has been slow economic growth and a structural budget imbalance for the past decade, debt has increased to 75% of GDP and debt service costs are now crowding out other spending priorities.

The main problem on the spending side of the budget is that a 5.5% increase has been agreed for public service salaries in 2025/26, over 1 percentage point higher than nominal growth last year and more than 2 percentage points higher than the current consumer price inflation rate.

In addition to adjustments for higher personnel costs, the 2025 Budget adds additional funding (by comparison with 2024 plans) for infrastructure investment, early retirement costs, education, health and other “frontline” services, employment programmes and modernisation of SARS and Home Affairs. Consolidated expenditure is projected to increase by 7.8% over the 2024/25 total.

With spending growth substantially exceeding nominal GDP growth, and no room for raising the budget deficit, a rise in the tax burden was an arithmetic inevitability.

On the 12th March, the minister tabled a revised budget, with the original spending plans largely unchanged but a rather different approach to raising additional revenue. The budget now goes through a parliamentary process governed by the Money Bills Amendment Procedure and Related Matters Act of 2009.

The Parliamentary budget process

The Constitution prescribes that legislation should provide for budget amendment procedures. The 2009 Act that gives effect to this was an initiative of parliament’s finance committee. It provides “fiscal responsibility” safeguards, both in the sequencing of budget approval or amendment processes and in the establishment of a Parliamentary Budget Office to provide expert advice. 

In brief, the Act provides for:

• Referral of the “fiscal framework” to the finance committees of the National Assembly and the National Council of Provinces, who must report within 16 days (or within a reasonable period),
• Consideration of the Division of Revenue Bill and Appropriation Bill by the appropriation committees after the fiscal framework has been approved by the National Assembly,
• Passing of a Division of Revenue Act within 35 days of approval of the fiscal framework,
• Approval or amendment of an Appropriation Bill only after a Division of Revenue Bill has been passed,
• Completion of the review of proposed appropriations by portfolio committees, their consideration by the appropriation committees and parliamentary approval with or without amendments or rejection of the appropriation bill within four months of the start of the financial year (i.e. by 1 August), and
• Amendment of revenue bills only if the revenue to be raised is consistent with the approved fiscal framework.

These procedures provide an orderly approach to the consideration of budget amendments, including requirements for consultation with “affected” ministers if amendments to appropriations are proposed. However, the Act does not clarify how political differences are to be resolved in the context of disagreement between the minister of finance and other members of cabinet, or between cabinet and parliament itself. Alongside the parliamentary process, a political bargain will have to be sought.

The focus falls initially on the “fiscal framework”. This is defined in the Act as follows:

“…the framework covering the period of the MTEF that gives effect to the national executive’s macro-economic policy and includes – 
(a)  estimates of all revenue proposals, budgetary and extra-budgetary specified separately…
(b)  estimates of all expenditure, budgetary and extra-budgetary specified separately…
(c)  estimates of borrowing…
(d)  estimates of interest and debt servicing charges, and 
(e)  an indication of the contingency reserve necessary for an appropriate response to emergencies or other temporary needs, and other factors based on similar objective criteria”

The intention appears to be that revenue “proposals,” such as an increase in VAT or changes to income tax, should be considered as part of the fiscal framework, whereas it is the aggregates of expenditure that are covered rather than details of the division of revenue or departmental appropriations (which come under later scrutiny). 

But this understanding of the “fiscal framework” is unfortunately not given formal expression in the treasury’s budget documentation. 

One difficulty is that the treasury’s “consolidated” budget framework includes projected expenditure by provinces, whereas the Money Bills Amendment Act envisages separate amendment processes to be followed by provincial legislatures. Provincial spending is mainly funded through transfers from the national budget, provinces have strictly curtailed borrowing powers, and expenditure on basic education, health services and other provincial functions clearly should be taken into account in considering macroeconomic and fiscal policy. So it seems sensible that this deviation from the letter of the Act should be accommodated in parliament’s review process.

But the requirement for “separate” specification of extra-budgetary revenue and expenditure raises more difficult challenges. 

The treasury includes 195 entities of various kinds in its consolidated budget estimates. They include social security funds, sector education and training authorities (SETAs) and the national skills fund, the national roads agency, PRASA, the national science councils, several water boards and a wide variety of regulatory and service delivery agencies. Their revenue includes user charges, statutory levies or taxes, administrative service fees and in many cases transfers or operating subsidies from national departmental votes. 

The Budget Review and its statistical annexure do not provide an aggregate estimate or separate specifications of all “extra-budgetary” funds and accounts, though the details can be inferred from revenue and expenditure tables for public entities that are included in the Estimates of National Expenditure. Parliament’s problem is that although these estimates are available in the documentation, they are not subject to statutory appropriation, and treasury control of “extra-budgetary” revenue and expenditure is haphazard. It turns out that some of the fastest growing spending estimates in this year’s budget are in these extra-budgetary categories.

On tax

The minister’s initial VAT proposal would have raised about R60 billion in additional revenue. It was accompanied by personal income tax (PIT) relief to (largely) offset the fiscal drag effects of inflation, and no adjustment to the fuel levy, equivalent to about R4 billion in real tax relief. The overall impact would have been an increase in the tax burden equivalent to nearly 1% of GDP, largely effected through a rise in consumer prices under conditions of slow growth in household disposable income.

In the revised tax proposals, VAT increases from 15% to 15.5% in May this year and by a further 0.5% in April 2026. The proposed personal income tax relief has been dropped, so that for the second year in a row there will be no change to the PIT thresholds and rebates or to medical tax credits. This will add about R20 billion to revenue in 2025/26. The revised VAT increase together with additional food items to be zero-rated is expected to yield an additional R11.5 billion in revenue. The fuel levy remains unadjusted, which is something of a surprise, taking into account health and climate change considerations and the comparative strength of the rand.

Personal income tax and VAT are the two largest sources of tax revenue. Whereas the VAT rate has been adjusted just once since 1993 – it was increased from 14% to 15% in 2018 – the effective PIT tax burden changes every year as a result of threshold and other adjustments (or not). These impact on households of these adjustments is perhaps less obvious than the impact of a VAT change, and so fiscal drag is sometimes referred to as a “stealth tax”.

Since 2014, the average rates of personal income tax have increased by about 3 percentage points for most taxpayers, calculated for constant price (CPI-adjusted) income cohorts. This has mainly been the result of below-inflation threshold adjustments, and the addition of a 45% bracket in 2016/17. The lower tax threshold (currently R95,750 per annum) has declined by about 20% in real terms. In effect, over the past decade, the personal income tax burden has increased by substantially more than the treasury’s proposed 2 percentage point increase in VAT – though the PIT adjustments have taken effect more gradually and do not affect lower-income households.

Taking a longer-term view, a rather different perspective emerges. In the early 1990s, in a context of slow growth and an unsustainable budget deficit, the effective PIT burden increased substantially: between 1989 and 1995, the top threshold remained unchanged at R80,000 and for several years there were no bracket adjustments despite inflation of 6-10% a year. But between 1996 and 2014, the accumulation of real tax relief amounted to around 10% of taxable income for taxpayers up to a taxable income level of around R1.5 million in 2023 prices. A taxable income of R750,000 in 2023 prices attracted tax of 37.5% in 1996, and 23.6% in 2014. 

These comparisons are potentially misleading. Since 1996, the tax base has broadened considerably, fringe benefits and capital gains are more effectively taxed, and SARS has strengthened its administrative capacity. The reduction in tax rates, particularly between 1999 and 2009, did not result in a significant loss of revenue as a percentage of GDP. It is apparent, on the other hand, that the increase in the PIT burden since 2014, through its effect on disposable household income, has in part been offset by reduced revenue from VAT and other consumption taxes and from corporate income.

This is likely to happen again in 2025/26. Consumer demand will be held back by depressed after-tax income of households, the projected GDP growth rate of 1.9% will not be achieved, and revenue will again fall short of the treasury’s budget estimates.

Consolidated expenditure

While it accounts for 90% of consolidated expenditure, only 14% of the national “main” budget goes to current payments for services, including compensation of employees, and capital spending. Debt service costs account for 18%, transfers to provinces and municipalities 41%, social grants 12% and other transfers and subsidies 15%. Goods and services purchased by national departments account for just 4% of the national budget. Perhaps there are savings to be found from reducing waste, fruitless expenditure or unproductive programmes of national departments, but this is not where the bulk of spending occurs.

The “front-line” health and education services funded by provinces are currently experiencing considerable financial stress. Insufficient funds for personnel, equipment and maintenance of facilities are in large measure a consequence of salary increases and staffing requirements that have exceeded available funds during recent fiscal consolidation years. The consolidated budget for the “social wage” continues to be prioritised in the treasury’s budget plans, but it has been over-stretched in recent years by additional allocations for student financial aid, the social relief of distress grant, early childhood development and new employment programmes.

Alongside strains in social spending programmes, government’s long-standing commitment to increase infrastructure investment in support of sustainable growth has not yet borne fruit. The budgeted public sector infrastructure spending plans again add up to about R1 trillion over the next three years, mainly by public entities and state-owned companies. But as in the past, actual spending is likely to fall short of these plans by 20-25%.

Perhaps unsurprisingly, the largest spending increases in this year’s budget are in extra-budgetary entities that are not directly controlled by either the treasury or parliamentary appropriations. In the treasury’s published “summary of the consolidated budget”, which perhaps approximates most closely to the “fiscal framework” required by the Money Bills Amendment Act, national main budget expenditure in 2025/26 increases, by comparison with the 2024 Budget forward estimate, by 3%. Additional spending attributed to provinces, social security funds and public entities increases by 16.5%.

These uncontrolled spending trends range across many agencies, amongst others:

• Board remuneration and executive salaries that often substantially exceed remuneration levels in the regulated public service,
• Diversion of funds to address budget programme shortfalls, for example in the use of National Skills Fund and SETA surpluses to finance student financial assistance,
• Allocation of social security funds for unregulated discretionary purposes, such as the use of UIF surpluses to bail out failing businesses and to supplement departmental personnel budgets,
• A substantial expansion of expenditure by the national roads agency, financed by drawing on cash reserves and an implausibly large increase in toll revenue estimates,
• Increased operating costs of PRASA, accommodated mainly by larger transfers from the main budget,
• Large increases in borrowing and expenditure by the extra-budgetary water resource agency, TCTA,
• Further growth in the unfunded liabilities of the Road Accident Fund, for which reform proposals proposed 23 years ago by the Satchwell Commission have not yet been implemented.

Whatever the merits of these and other spending activities of off-budget funds and public entities, it seems clear that stronger executive and parliamentary oversight of their spending plans is needed in future. 

On debt

The consolidated budget deficit was 5% of GDP in 2024/25, and the wider public sector borrowing requirement was 5.3%. For the year ahead, the treasury anticipates a public sector borrowing requirement of 5.9% of GDP. Eskom and Transnet are still financially stressed. Debt is growing faster than the economy, at interest rates substantially higher than inflation.

The treasury kept gross loan debt of the national government to 76.1% of GDP by the end of 2025/26, up from 74.1% at the end of 2024/25, by drawing R100 billion into the fiscus from the accumulated balance of the Gold and Foreign Exchange Contingency Reserve Account. There will be a further R25 billion drawdown in 2025/26 and the Eskom debt-settlement commitment for the year has been reduced by R30 billion. The proposed tax increase also contributes to moderating the rise in debt, nonetheless projected to increase to 76.2% of GDP by the end of 2025/26.

Special measures have been needed to contain the debt increase last year and for the year ahead. But interest rates remain elevated, debt service costs continue to rise as a share of revenue and the commercial ratings agencies still classify South Africa’s sovereign debt as below investment-grade. 

Budget reform and institutional change

Macro-fiscally, South Africa’s growth outlook has not yet turned, the scope for further tax increases appears to be limited and the debt-GDP ratio has not yet reached an inflection point.

For many years, a “comprehensive expenditure review” has been under discussion, but there is no evidence yet of decisive measures to curtail unproductive activities or reverse the growth in service delivery and programme commitments. Although over 200 programme-targeted expenditure reviews have been undertaken, their focus has been on improving programme design rather than identifying savings.

In his budget speech on 12 March this year, Minister Godongwana emphasised that a budget reform process is underway, provincial and local government grant frameworks are under review and the treasury is committed to eliminating waste and undertaking systematic expenditure reviews, under guidance of the Presidency.

It seems clear that the current political turmoil over revenue proposals in the 2025 Budget must lead to new approaches to expenditure planning and review. This must include stronger oversight of extra-budgetary accounts and funds, a more disciplined approach to public sector salary determination under conditions of fiscal constraint, reconsideration of the balance between public and private sector responsibilities and improved cost recovery in several sectors.

Better analysis and public engagement on trade-offs and priorities might contribute to better budgeting and more sustainable public finances. The Parliamentary Budget Office and the Financial and Fiscal Commission have statutory responsibilities for fiscal research and advice and might assist parliament in exercising its money bills amendment powers. Both the treasury’s budget reform initiatives and parliamentary processes might involve constructive institutional change.

But whether we are at an inflection point that might lead to decisive action on spending priorities and commitments is fundamentally a political question.