# Emergency Ageing Austerity Identifier: GSX:POP-005 Series: Population Futures Version: 1.0 Status: published Schema: GreyResearchObject 1.2 ## Primary question Which policy packages for closing an ageing-related fiscal deficit are examined for lower welfare loss? ## Summary Policy packages for closing an ageing-related fiscal deficit with lower welfare loss. The complete source manuscript is available as Markdown. Published claim objects are a claim-led extraction, not an exhaustive list of every statement in the paper. Read each claim with its assumptions and main limitation; the catalog summary is not a finding. ## Conditional answer Among five packages that each close a hypothetical R150 billion annual ageing gap, the balanced transition has the lowest modelled welfare-loss index, 64.5. The low-visible-pain package, which leans on inflation, diffuse cuts, and continuing borrowing, scores 107.3 and leaves about R720 billion of emergency debt after ten years. - Evidence status: modelled - Depends on: GSX:POP-005:A01, GSX:POP-005:A02 - Derived from: GSX:POP-005:R01, GSX:POP-005:R02 - Sensitivity: high - Main limitation: The R150 billion gap is a stress test, not a forecast of today's deficit. The welfare scores are internal model indexes, not percentages and not estimates of national wellbeing. The manuscript does not declare any package the single correct answer. ## Related questions - How can an ageing-related fiscal deficit be closed? - What is the welfare comparison among emergency responses to demographic fiscal pressure? ## Epistemic status - Study type: Exploratory simulation - Evidence status: Modelled - Peer review: None - External critique: None recorded - AI assistance: GreyScienx manuscripts are prepared with substantial AI assistance in research, drafting, and computation. This record is not peer-reviewed. The published manuscript does not itemise AI participation by activity. - Record date: 2026-09-28. Date of this machine record. It is not a separately catalogued manuscript issue date. ## Topics - [fiscal-policy](https://greyscienx.com/topics/fiscal-policy) - [retirement](https://greyscienx.com/topics/retirement) - [population-decline](https://greyscienx.com/topics/population-decline) - [south-africa](https://greyscienx.com/topics/south-africa) ## Geographies - South Africa ## Assumptions - [A01](https://greyscienx.com/research/population-futures/emergency-ageing-austerity/assumptions/A01) Stress-test gap. The study imposes a hypothetical R150 billion annual ageing-related fiscal gap in constant 2026 rand. The figure is a stress-test target, about 5.6 percent of the R2.67 trillion in spending cited from the 2026 Budget Speech. It is not a forecast of the current deficit. - [A02](https://greyscienx.com/research/population-futures/emergency-ageing-austerity/assumptions/A02) Thirteen capped levers. Thirteen policy levers are each capped at a visible but limited level. Their mature annual yields are scenario capacities assigned by the model, not official costings. ## Model - [M01](https://greyscienx.com/research/population-futures/emergency-ageing-austerity/models/M01) Ageing-austerity package comparison. The model asks capped levers to close one annual gap and scores five packages on an internal welfare-loss index, political resistance, year-one borrowing, and debt after ten years. Lower index values are better. The scores are not percentages of national wellbeing. ## Results - [R01](https://greyscienx.com/research/population-futures/emergency-ageing-austerity/results/R01) All five selected packages close R150 billion once fully implemented. The balanced transition has the lowest welfare-loss index, 64.5. Low visible pain scores 107.3 and leaves about R720 billion of emergency debt after ten years. (64.5 internal welfare-loss index, balanced transition; Five packages, each closing the R150 billion gap). - [R02](https://greyscienx.com/research/population-futures/emergency-ageing-austerity/results/R02) The balanced transition uses R57 billion of targeted pension measures, R24 billion of additional pre-funding, R26 billion of worker tax, R22 billion from a phased retirement-age change, R6 billion of weaker indexation, and R15 billion of continuing borrowing. It avoids healthcare and non-age service cuts. (57 billion constant 2026 rand of targeted pension measures; Balanced-transition package). ## Sources - GSX:POP-005:PDF [Emergency Ageing Austerity](https://greyscienx.com/research/population-futures/emergency-ageing-austerity.pdf) — Published manuscript used for this semantic-twin extraction. External references remain represented in the manuscript unless separately registered. - GSX:POP-005:SRC01 [2026 Budget Speech](https://www.treasury.gov.za/documents/national%20budget/2026/speech/speech.pdf) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim. - GSX:POP-005:SRC02 [Statistics South Africa](https://www.statssa.gov.za/?p=19307) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim. - GSX:POP-005:SRC03 [National Treasury's 2026 Estimates of National Expenditure](https://www.treasury.gov.za/documents/national%20budget/2026/ene/FullENE.pdf) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim. - GSX:POP-005:SRC04 [National Treasury](https://www.treasury.gov.za/documents/National%20Budget/2026/review/Chapter%204.pdf) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim. - GSX:POP-005:SRC05 [SARS](https://www.sars.gov.za/latest-news/retirement-fund-contribution-deductions-section-11f2a/) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim. - GSX:POP-005:SRC06 [South African Reserve Bank](https://www.resbank.co.za/content/dam/sarb/publications/reviews/monetary-policy-review/2026/april.pdf) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim. - GSX:POP-005:SRC07 [National Treasury's 2026 Budget Review](https://www.treasury.gov.za/documents/National%20Budget/2026/review/Chapter%203.pdf) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim. ## Manuscript - [Emergency Ageing Austerity (PDF, 15 pages)](https://greyscienx.com/research/population-futures/emergency-ageing-austerity.pdf) - Datasets: https://greyscienx.com/research/population-futures/emergency-ageing-austerity/data - Reproducibility: https://greyscienx.com/research/population-futures/emergency-ageing-austerity/reproduce ## Related studies - [When Retirement Becomes Impossible](https://greyscienx.com/research/population-futures/when-retirement-becomes-impossible) (GSX:POP-001) - [Today’s Unemployment Is Tomorrow’s Pension Crisis](https://greyscienx.com/research/population-futures/todays-unemployment-is-tomorrows-pension-crisis) (GSX:POP-002) - [Does Raising the Retirement Age Actually Work?](https://greyscienx.com/research/population-futures/does-raising-the-retirement-age-actually-work) (GSX:POP-003) - [The Politics of an Ageing Electorate](https://greyscienx.com/research/population-futures/the-politics-of-an-ageing-electorate) (GSX:POP-004) - [The Price of Another Child](https://greyscienx.com/research/population-futures/the-price-of-another-child) (GSX:POP-006) - [The Hundred-Year Life](https://greyscienx.com/research/population-futures/the-hundred-year-life) (GSX:POP-007) - [Inheritance After Retirement](https://greyscienx.com/research/population-futures/inheritance-after-retirement) (GSX:POP-008) - [How to Shrink a Country Without Breaking It](https://greyscienx.com/research/population-futures/how-to-shrink-a-country-without-breaking-it) (GSX:POP-009) - [The Scarce-Worker Economy](https://greyscienx.com/research/population-futures/the-scarce-worker-economy) (GSX:POP-010) - [The Population System](https://greyscienx.com/research/population-futures/the-population-system) (GSX:POP-011) - [Extreme Fiscal Pressure](https://greyscienx.com/research/population-futures/extreme-fiscal-pressure) (GSX:POP-012) ## Machine-readable record - JSON: https://greyscienx.com/api/v1/research/emergency-ageing-austerity - Full-text manuscript API: https://greyscienx.com/api/v1/research/emergency-ageing-austerity/manuscript - Canonical page: https://greyscienx.com/research/population-futures/emergency-ageing-austerity - Citation: GreyScienx. Emergency Ageing Austerity. GSX:POP-005. Version 1.0. ## Claim-level record ### GSX:POP-005:C01 Among five packages that each close a hypothetical R150 billion annual ageing gap, the balanced transition has the lowest modelled welfare-loss index, 64.5. The low-visible-pain package, which leans on inflation, diffuse cuts, and continuing borrowing, scores 107.3 and leaves about R720 billion of emergency debt after ten years. - Status: conditional - Evidence status: modelled - Depends on: GSX:POP-005:A01, GSX:POP-005:A02 - Result: GSX:POP-005:R01 - Sensitivity: high - Main limitation: The R150 billion gap is a stress test, not a forecast of today's deficit. The welfare scores are internal model indexes, not percentages and not estimates of national wellbeing. The manuscript does not declare any package the single correct answer. - Page: https://greyscienx.com/research/population-futures/emergency-ageing-austerity/claims/C01 ### GSX:POP-005:C02 The balanced transition in the model closes the gap with R57 billion of targeted pension measures, R24 billion of additional pre-funding, R26 billion of worker tax, R22 billion from a phased retirement-age change, R6 billion of weaker indexation, and R15 billion of continuing borrowing. It does not cut healthcare or non-age services. - Status: conditional - Evidence status: modelled - Depends on: GSX:POP-005:A01, GSX:POP-005:A02 - Result: GSX:POP-005:R01 - Sensitivity: moderate - Main limitation: These are the manuscript's assigned scenario yields for one package. They are not an official budget costing, and a different cap on any lever changes the mix. - Page: https://greyscienx.com/research/population-futures/emergency-ageing-austerity/claims/C02 2 assumptions, 2 results, and 1 model descriptions are included in the JSON record. ## Full manuscript Source checksum: 913ccaac3c2644070bf1aaae41fcdfc596040ca71039d8e8046e90b44b5d3440 The text below is the complete Markdown source used to produce the manuscript, preserved separately from the semantic claim layer above. --- # Emergency Ageing Austerity ## Closing a fiscal gap without choosing a sacrificial generation # Part I: The finding There is no painless ageing-austerity package. There are only different choices about who pays, when they pay and whether the adjustment protects future productive capacity. This study imposes a hypothetical R150 billion annual ageing-related fiscal gap in constant 2026 rand. It then asks thirteen policy levers to close it: worker and consumption taxes, pension contributions, a later retirement age, benefit reductions, means testing, solidarity taxes, pension-wealth taxation, weaker indexation, inflation, healthcare cuts, non-age spending cuts and borrowing. Five packages sit near different points of the political and welfare trade-off. All close R150 billion once fully implemented. None is declared the single correct answer. | Package | Welfare-loss index | Political resistance | Year-one borrowing | Debt after 10 years | |---|---:|---:|---:|---:| | Balanced transition | 64.5 | 98.2 | R67bn | R394bn | | Targeted solidarity | 68.4 | 99.1 | R43bn | R272bn | | Work-and-save transition | 66.0 | 96.2 | R52bn | R292bn | | Protect basic benefits | 69.0 | 94.1 | R52bn | R384bn | | Low visible pain | 107.3 | 53.5 | R75bn | R720bn | Lower index values are better. These are internal model scores, not percentages and not estimates of national wellbeing. The balanced transition has the lowest total welfare loss among the five selected packages. It uses R57 billion of targeted pension measures, R24 billion of additional pre-funding, R26 billion of worker tax, R22 billion from a phased retirement-age change, R6 billion of weaker indexation and R15 billion of continuing borrowing. It avoids healthcare and non-age service cuts. The targeted-solidarity package is slightly more costly in the welfare index but creates the smallest ten-year debt bridge because more of its adjustment works immediately. It leans more heavily on current taxes and affluent pensioners while avoiding a higher retirement age and broad benefit restraint. The politically easiest package is the worst social bargain. "Low visible pain" avoids large visible tax and benefit changes by using inflation, diffuse spending cuts and R52.5 billion of continuing annual borrowing. Its political-resistance score is almost half that of the balanced package, but its welfare loss is two-thirds higher and its emergency debt reaches about R720 billion after ten years. > The efficient answer is a portfolio: target capacity to pay, preserve basic benefits and high-return services, announce slow reforms early, and use borrowing only as a bridge. Political invisibility is not economic efficiency. # Part II: What is an ageing emergency? An ageing-related deficit is the part of the public shortfall created when age-linked spending and healthcare grow faster than the tax base that supports them. It is not identical to the current budget deficit. A government can have an ageing gap while its headline budget is balanced, or a large budget deficit for reasons unrelated to ageing. The R150 billion figure is a stress-test target, not a forecast. It is large enough to require genuine trade-offs but smaller than South Africa's present annual debt-service bill. It is roughly 5.6 percent of the R2.67 trillion in spending announced in the [2026 Budget Speech](https://www.treasury.gov.za/documents/national%20budget/2026/speech/speech.pdf). The pressure is plausible in direction. [Statistics South Africa](https://www.statssa.gov.za/?p=19307) reports that the number of people aged 60 or older rose from 3.6 million in 2002 to 6.6 million in 2025. Grants were the main income source for 60.1 percent of older-person-headed households in 2024, while 68.5 percent of older people relied on public health facilities. [National Treasury's 2026 Estimates of National Expenditure](https://www.treasury.gov.za/documents/national%20budget/2026/ene/FullENE.pdf) reports about 4.2 million old-age grant beneficiaries in 2025/26 and a target of 4.3 million in 2026/27. The Budget Speech sets the old-age grant at R2,400 a month from April 2026 and allocates R292.8 billion to all social grants. The emergency begins when ordinary parameter changes are postponed for so long that government must adjust quickly. A one-year delay may appear compassionate, but repeated delay narrows the menu. Gradual contribution changes and retirement-age transitions no longer arrive in time, leaving abrupt taxes, broad benefit cuts, service reductions, inflation or borrowing. # Part III: The thirteen levers Each policy is capped at a deliberately visible but not unlimited level. The fiscal yield is the maximum mature annual contribution assigned by the model. These are scenario capacities, not official costings. | Intervention | Modelled maximum | Mature annual yield | First-year share | |---|---:|---:|---:| | Worker income-tax surcharge | +3 points | R100bn | 100% | | Broad consumption-tax increase | +2 points | R55bn | 100% | | Higher retirement contributions | +4 points | R24bn | 10% | | Raise pension and grant age | +5 years | R55bn | 18% | | Reduce pension benefits | -10% | R36bn | 100% | | Means-test affluent pensioners | Top quarter targeted | R30bn | 72% | | Solidarity tax on high pensions | +5 points | R18bn | 100% | | Tax private pension wealth | 0.5% annually | R25bn | 85% | | Index benefits below inflation | 3 points below CPI | R28bn | 32% | | Inflation surprise | +5 points | R42bn | 100% | | Reduce age-related healthcare | -10% | R35bn | 100% | | Cut non-age government spending | -5% | R80bn | 100% | | Additional annual borrowing | R75bn | R75bn | 100% | The contribution lever measures reduced future public dependence once higher private saving has matured. It does not pretend that pension contributions are immediate tax revenue. The retirement-age and indexation levers also phase in. Borrowing closes a cash gap today but never closes the structural gap; it moves the bill forward and adds interest. The model divides every lever into five tranches. Early tranches have lower welfare and political costs. Later tranches become progressively more painful. This captures a basic policy reality without equations: the first rand from a broad tax base is usually easier than the last, and moderate reform is less disruptive than forcing one instrument to its limit. # Part IV: How the simulator judges a package The simulator applies four tests. First, a mature package must close the full R150 billion annual gap. Second, it records how quickly the structural measures arrive and fills the transition shortfall with borrowing. Third, it assigns the welfare loss to workers, pensioners, children and future taxpayers. Fourth, it gives each intervention a political-resistance cost based on visibility, concentration and likely organisation. | Group | Typical burdens in the model | Why the burden matters | |---|---|---| | Workers | Income tax, consumption tax, contributions and some inflation | Lower disposable income can reduce consumption and labour supply | | Pensioners | Retirement-age changes, benefits, indexation, health and targeted taxes | Fixed incomes and short adjustment horizons raise vulnerability | | Children | Consumption tax and cuts to education or household services | Losses can compound across an entire lifetime | | Future taxpayers | Borrowing and foregone productive investment | They pay without voting in the decision | The welfare weights distinguish a rand transferred from a high pension income from a rand removed from a basic grant or essential treatment. Means testing and solidarity taxes therefore impose less welfare loss per rand raised than broad benefit or healthcare cuts. Higher contributions also receive a relatively low cost because most of the contribution becomes the worker's asset, although current liquidity still falls. The political score is deliberately different. Borrowing, inflation and diffuse spending restraint initially face less resistance than a visible pension cut or retirement-age increase. The model does not reward that invisibility as welfare. It exposes the divergence. The optimisation is not a moral calculator. Its weights are declared assumptions. It omits administrative gaming, tax avoidance, labour-market responses, party competition and legal constraints. The ranking is useful only to the extent that the reader agrees with the relative burdens. # Part V: Capacity and timing No single bounded intervention closes the R150 billion gap. A maximum worker income-tax surcharge reaches R100 billion in the model, non-age cuts R80 billion and borrowing R75 billion. Every other lever is smaller. ![Intervention capacity](assets/intervention-capacity.png) *Figure 1. Grey bars show mature annual fiscal capacity; coral bars show the first-year effect. Higher contributions, retirement ages and indexation changes need time.* The timing difference changes emergency policy. Taxes, broad benefit cuts, healthcare cuts and borrowing can affect the first budget immediately. A later retirement age applies only to newly affected cohorts if it is phased fairly. Higher contributions reduce disposable income now but reduce public dependence much later. Lower indexation begins gently and compounds. This creates a two-part problem. Government needs a structural package that eventually closes R150 billion and a financing plan for the years before it matures. Treating these as one decision creates two common errors: counting future pension savings as cash available today, or treating temporary borrowing as a permanent solution. # Part VI: Five efficient combinations The five packages are constrained optimisations. Each protects a different value and then selects the least costly available tranches under that constraint. ![Package composition](assets/package-composition.png) *Figure 2. Mature annual contribution to the R150 billion gap. "Borrowing" is the annual portion that remains unclosed structurally even after the package matures.* | Package | Main design choice | What it avoids | |---|---|---| | Balanced transition | Diversified targeted measures, pre-funding, modest tax and age reform | Health and non-age service cuts | | Targeted solidarity | Affluent-pension measures plus current tax | Retirement-age and broad benefit reductions | | Work-and-save transition | Contributions, limited age change and targeted pension revenue | Large service cuts and inflation | | Protect basic benefits | Taxes, pre-funding and affluent-pension measures | Basic benefit, indexation and healthcare cuts | | Low visible pain | Borrowing, inflation and diffuse cuts | Large visible current tax or pension reform | The balanced package uses the full four-point contribution increase, all three targeted pension instruments and roughly a two-year age movement. Targeted solidarity replaces age and broad-benefit restraint with R52.8 billion of current tax and R62 billion from affluent pensioners, giving it the smallest transition debt. Protect basic benefits avoids benefit, indexation and healthcare cuts, but still asks affluent pensioners to contribute. # Part VII: Who pays? Closing the same fiscal gap does not produce the same social outcome. The balanced package assigns 45 percent of its modelled welfare loss to pensioners, 33 percent to workers, 5 percent to children and 16 percent to future taxpayers. Much of the pensioner share is deliberately concentrated on affluent households. ![Distributional incidence](assets/distributional-incidence.png) *Figure 3. Shares refer to weighted welfare loss, not rands of tax or spending. A smaller total loss can have a larger percentage assigned to one group.* Targeted solidarity shifts 44 percent of welfare loss to workers because it uses more income and consumption tax, while pensioners carry 36 percent. The work-and-save package is similar, but slightly less dependent on current taxation and more dependent on later work. Protecting basic benefits moves 21 percent of loss to future taxpayers, mainly because R21 billion of annual borrowing remains after maturity. Low visible pain moves 41 percent of welfare loss to the future and 18 percent to children through borrowing, inflation and non-age cuts. The figure also reveals an important limitation of generational categories. A worker may be a future pensioner. A pensioner may support children in the same household. A child benefits when an old-age grant stabilises household food spending. The incidence categories identify the first-round bearer, not the complete household network. # Part VIII: The efficient frontier Welfare efficiency and political ease do not select the same policies. The frontier traces the best available package as the optimiser gives progressively more weight to political resistance. ![Efficiency frontier](assets/efficiency-frontier.png) *Figure 4. Lower-left is preferred. Named packages include additional distributional protections, so some sit above the unconstrained frontier.* Near the right-hand end, the optimiser minimises total welfare loss. It accepts visible taxes, targeting and pre-funding. Moving left reduces political resistance by substituting borrowing, inflation and diffuse cuts. The welfare curve rises slowly at first, then sharply as the apparently easy instruments dominate. The balanced, work-and-save and targeted packages cluster near a welfare-loss index of 65 to 68 but have resistance scores around 96 to 99. This is the central political challenge: the packages that preserve services and limit debt require governments to announce who pays. Low visible pain has a resistance score of only 53.5, yet its welfare index rises to 107.3. It is not inefficient because the model forgot politics. It is inefficient because it optimises politics at the expense of unrepresented or weakly represented groups. There is no unique point where the frontier becomes morally correct. A government choosing among these packages must state its distributional priorities rather than hiding them inside technical language. # Part IX: Taxes and pre-funding The tax options differ in both incidence and credibility. A worker income-tax surcharge is progressive if it follows the existing schedule, but the tax base is relatively narrow. National Treasury estimates final personal income tax of R844.8 billion for 2026/27 and about 8.34 million individuals above the tax threshold. A R100 billion maximum model yield is therefore substantial, roughly 12 percent of projected personal income tax, and should be read as a high stress-test bound. [National Treasury](https://www.treasury.gov.za/documents/National%20Budget/2026/review/Chapter%204.pdf) A broad consumption-tax increase raises revenue quickly and reaches informal as well as formal earners. That breadth is its strength and its distributional weakness. Lower-income households consume more of their income, so zero-rating, rebates or compensating grants matter. The model assigns a quarter of its welfare burden to children through household budgets. Higher pension contributions are different. They reduce current take-home pay but build private assets. [SARS](https://www.sars.gov.za/latest-news/retirement-fund-contribution-deductions-section-11f2a/) states that deductible retirement-fund contributions are generally limited to 27.5 percent of remuneration or taxable income, subject to the statutory cap. The model's four-point increase should fit inside that broad legal context for many workers, but not for everyone. Pre-funding is economically attractive and fiscally slow. Its full R24 billion modelled relief appears only after ten years. It therefore belongs in an early ageing strategy, not as the sole response once the emergency has arrived. # Part X: Pension policy is not one lever Pension restraint can be targeted by wealth, spread across all benefits, delayed through indexation or imposed through a later age. These choices are not equivalent. Means testing and a solidarity tax preserve the benefit floor while asking more from people with private capacity. They score well on welfare in this model, but they require reliable income and asset information. Poor administration can create exclusion errors, avoidance, arbitrary thresholds and high compliance costs. A private pension-wealth tax broadens the target but creates valuation and liquidity problems. It may also contradict earlier policy promises made to encourage long-term saving. The model therefore gives it a higher political cost than a solidarity tax on pension income. A broad ten-percent benefit reduction is immediate and administratively simple. It is also blunt. It removes the same percentage from a low-income recipient and an affluent pensioner unless the system contains a floor. For that reason, none of the four welfare-focused packages uses more than a small fraction of the available broad cut. Reduced indexation is slower and less visible. Three percentage points below inflation eventually closes R28 billion in the scenario, but only about one-third appears in the first year. Its cumulative character can turn a modest initial adjustment into a large real-income loss for the oldest pensioners. Raising the pension and grant age creates fiscal savings only if affected people can work or receive suitable bridge support. The balanced package uses roughly a two-year shift, while targeted solidarity uses none. This reflects the earlier GreyScienx finding that a legal retirement age does not create a job. # Part XI: Inflation, services and debt Inflation is the most deceptive benefit cut. It reduces the real value of nominal pensions, wages and cash balances without a line item called "austerity." It also changes interest rates, exchange rates, contracts and expectations across the economy. South Africa's monetary framework now targets 3 percent inflation with a one-point tolerance band. The [South African Reserve Bank](https://www.resbank.co.za/content/dam/sarb/publications/reviews/monetary-policy-review/2026/april.pdf) describes price stability as protection for purchasing power and a foundation for investment and balanced growth. The model's five-point inflation surprise is therefore an adverse scenario, not a policy recommendation. Its apparent political ease ignores later monetary tightening and credibility loss. Healthcare cuts are visible in suffering rather than in a pension statement. Their maximum model yield is R35 billion, but they carry the highest welfare cost per rand. That judgement is reinforced by the high public-health reliance of older South Africans reported by Statistics South Africa. Non-age cuts can raise more, up to R80 billion in the model, because the spending base is broad. But "non-age" does not mean unnecessary. It includes education, housing, transport, safety and economic infrastructure. Reducing it can protect current benefits by weakening the future tax base. Borrowing is necessary during transition and dangerous as a permanent residual. [National Treasury's 2026 Budget Review](https://www.treasury.gov.za/documents/National%20Budget/2026/review/Chapter%203.pdf) projects debt-service costs at 20.8 percent of main-budget revenue in 2026/27, after 21.3 percent in 2025/26. The starting fiscal system therefore has limited room for an open-ended ageing liability. ![Transition borrowing](assets/transition-borrowing.png) *Figure 5. Annual borrowing covers the difference between the R150 billion gap and structural measures as they mature. Debt compounds at a modelled 3 percent real rate.* Ten-year debt ranges from R272 billion under targeted solidarity to R720 billion under low visible pain; the work-and-save, basic-benefit and balanced packages lie between R292 billion and R394 billion. # Part XII: The policy answer The study does not select one package because the ranking depends on values and timing. Choose the balanced transition when the priority is the lowest total welfare loss and government can credibly phase reform. Choose targeted solidarity when the emergency is immediate, data systems can distinguish affluent pensioners and limiting transition debt matters. Choose work-and-save when the labour market can support later work and households can absorb higher contributions. Choose protection of basic benefits when poverty and healthcare risks dominate, accepting more tax and borrowing. Reject low visible pain unless a genuine temporary shock makes short borrowing unavoidable and a funded exit is legislated at the same time. | Decision rule | Preferred direction | Main safeguard | |---|---|---| | Minimise total welfare loss | Balanced mix of targeting, saving, tax and gradual age reform | Protect low-income pensioners and bridge workers | | Minimise transition debt | Targeted solidarity plus current tax | Strong administration and anti-avoidance | | Protect basic benefits | Tax affluent pensions and workers, pre-fund, borrow modestly | Debt ceiling and scheduled review | | Protect children and growth | Avoid health, education and infrastructure cuts | Explicit expenditure floors | | Reduce visible opposition | Borrowing and hidden real cuts | Recognise that this is politically easy, not socially efficient | Six design principles survive every package. 1. Publish the full incidence: show workers, pensioners, children and future taxpayers separately. 2. Separate the first-year cash plan from the mature structural plan. 3. Use capacity-to-pay instruments before basic benefits and essential healthcare. 4. Preserve education, infrastructure and preventative health because austerity can create the next fiscal crisis. 5. Put automatic expiry dates on emergency taxes and explicit review dates on benefit rules. 6. Cap continuing borrowing and disclose the interest-adjusted burden inherited by future taxpayers. > Efficient austerity is not the package with the fewest protests this year. It is the package that closes the gap while doing the least lifetime damage and preserving political consent across generations. The fiscal arithmetic can identify feasible combinations. It cannot decide how much hardship society owes each group. That decision remains political. The model's contribution is to make the exchange explicit: every protected constituency creates a payer elsewhere, and every delayed reform creates a future taxpayer who was absent from the vote. # Notes: Scope, limitations and sources This is an armchair scenario model, not an actuarial valuation, official fiscal projection, tax forecast or policy recommendation. All values are constant 2026 rand. The annual ageing-related gap is fixed at R150 billion. It does not grow with GDP, demographics or healthcare prices. The model closes the mature annual gap exactly, but slow interventions require transition borrowing. Each intervention has a modelled maximum fiscal capacity, first-year realisation share, maturity period, welfare cost, political-resistance cost, four-group incidence and vulnerable-pensioner exposure. Each lever is divided into five tranches with rising marginal costs. A linear optimiser selects tranches subject to each named package's policy caps. The efficient frontier varies the weight placed on political resistance. The welfare-loss index divides weighted modelled loss by the R150 billion gap and multiplies by 100. It is an ordinal comparison inside this model. It is not a percentage of GDP, happiness or income. The political-resistance index is constructed similarly and is not an opinion-poll estimate. Group shares distribute weighted welfare loss, not fiscal cash. The transition path assumes a 3 percent real interest rate. Income taxes, consumption taxes, broad benefit cuts, inflation, healthcare cuts, non-age cuts and borrowing have their full fiscal effect in year one. Means testing and pension-wealth taxation mature over two years, a retirement-age change over six, indexation over five and higher contributions over ten. The contribution channel is long-run relief from reduced dependence, not revenue paid to the budget. The retirement-age channel combines delayed grants or pensions and continued tax contribution, but the model does not reproduce occupation-specific employment or disability. The inflation channel omits the full monetary-policy and balance-sheet response. Borrowing is counted as cash financing but not structural closure. The model omits behavioural tax responses, avoidance, labour supply, unemployment, firm investment, household transfers, pension-fund balance sheets, public-service pension rules, medical inflation, provincial budgets, legal rights, administration, take-up and distribution within the four broad groups. Affluent-pension measures are assumed to be administratively possible. In practice, data quality and avoidance could materially reduce yield. Primary context comes from [Statistics South Africa's older-population report](https://www.statssa.gov.za/?p=19307), the [2026 Budget Speech](https://www.treasury.gov.za/documents/national%20budget/2026/speech/speech.pdf), the [2026 Estimates of National Expenditure](https://www.treasury.gov.za/documents/national%20budget/2026/ene/FullENE.pdf), the [2026 Budget Review fiscal-policy chapter](https://www.treasury.gov.za/documents/National%20Budget/2026/review/Chapter%203.pdf), its [tax-policy chapter](https://www.treasury.gov.za/documents/National%20Budget/2026/review/Chapter%204.pdf), the [SARS retirement-contribution guidance](https://www.sars.gov.za/latest-news/retirement-fund-contribution-deductions-section-11f2a/) and the [South African Reserve Bank's April 2026 Monetary Policy Review](https://www.resbank.co.za/content/dam/sarb/publications/reviews/monetary-policy-review/2026/april.pdf).