Retirement math: who pays for pensions as populations age?

Retirement income can come from public pensions, workplace arrangements and personal assets. Those sources have different funding rules. Population ageing can put pressure on some promises, but it does not establish that everyone will be unable to retire.

The useful questions are who supplies the income, what can change, and how much your own plan depends on each assumption. Source observations and projections below were checked October 4, 2026. Personal examples use hypothetical currency units.

Distinguish people from contributors

The OECD’s demographic measure counts people aged 65 and over per 100 people aged 20–64. It does not count employed workers, taxpayers or pension recipients. Its 2025 report gives 32.6 for the OECD in 2024 and projects 55.2 in 2054 using UN medium-variant population projections. Taking the reciprocal gives about 3.07 working-age people per older person in 2024 and 1.81 in the projection.

Source: OECD demographic table 6.2. These ratios do not establish any country’s benefit amount. Employment, wages, productivity, migration, tax funding and pension rules also matter.

Follow the money rather than the label

In a pay-as-you-go arrangement, current contributions or taxes fund current benefits. Reserves can supplement that flow. In a funded arrangement, accumulated assets and their investment outcomes support payments. Countries and workplace schemes can combine mechanisms. Read retirement account basics for the distinction between an income promise and an invested account balance.

In a deliberately simplified example, 100 contributors each paying 10 units provide 1,000. Shared among 25 recipients, that funds 40 per recipient. With 40 recipients and the same inflow, it funds 25. Maintaining 40 requires total funding of 1,600. This excludes administration, other taxes, reserves and investment income: it explains a constraint, not an actual pension formula.

Possible responses include additional revenue, different benefit formulas or eligibility ages, other public funding and changes in employment or productivity. Personal saving is another income source; it does not automatically fill every household’s gap.

What reserve depletion means in the United States

The 2026 Social Security Trustees Report projects, under its intermediate assumptions, that Old-Age and Survivors Insurance reserves will be depleted in the fourth quarter of 2032. Projected continuing income would then cover 78% of scheduled benefits. This is a projected financing shortfall, not zero benefits.

The combined old-age, survivors and disability calculation is different: projected depletion in the third quarter of 2034, with 83% payable then. Keep funds, dates and percentages together. These projections are not a legislated reduction schedule or a guarantee about future reforms.

The same report’s contributor measure gives about 2.6 covered workers per OASDI beneficiary in 2025, projected to decline to about 1.9 by 2075. This is not the OECD population ratio: the numerator and denominator describe programme participants.

Country rules change the planning question

The OECD retirement-age comparison assumes labour-market entry at 22 in 2024 and an uninterrupted career under current legislation. It gives future average normal retirement ages of 66.4 for men and 65.9 for women, versus 64.7 and 63.9 for retirement in 2024. These are modelled comparisons, not every reader’s entitlement date.

• United Kingdom: the official State Pension timetable phases the increase from 66 to 67 during 2026–2028. Date of birth determines the step. This is a State Pension rule, not a universal workplace-scheme access age.
• China: the September 2024 legislative decision gradually raises statutory retirement ages over 15 years starting in 2025: men from 60 to 63, women in the two specified worker categories from 55 to 58 and from 50 to 55. The final ages do not apply to everybody immediately.
• Japan: the 2024 actuarial valuation, pages 25–27, describes modified indexation that adjusts benefit growth for demographic pressures. Outcomes vary with economic assumptions. This is more specific than saying pensions will simply disappear.

Use your actual scheme’s current statement for eligibility, contribution record, projected benefits, inflation adjustment and access restrictions. National replacement-rate percentages are not interchangeable personal forecasts.

Confidence is not a measured retirement gap

In the US Federal Reserve’s 2025 household survey, 35% of non-retired adults said their retirement saving was on track. This is a self-assessment in one country. It does not prove that the remaining 65% cannot retire or that respondents who feel on track have sufficient assets.

A personal gap needs actual expected income and costs. Assumed spending of 2,000 units monthly against assumed reliable pension receipts of 1,400 leaves 600 to fund elsewhere. Compare both on the same tax and purchasing-power basis. A monthly gap is not yet a required portfolio balance: duration, uncertain returns, withdrawals, longevity and other resources still matter.

Reproduce the savings calculation

Suppose you contribute 200 units at each month-end until age 67, starting with no invested balance. Assume a constant 5% annual effective real return, after inflation, before omitted fees and taxes. The equivalent monthly rate is i = 1.05^(1/12) − 1. For n months, the modelled balance is 200 × ((1 + i)^n − 1) / i.

The target column divides 100,000 by the same annuity factor. Contributions are constant in purchasing-power units: maintaining them when prices rise means increasing the nominal contribution. End values are in those same constant units.

Earlier starts include both more contributions and more compounding time. Contribution size matters directly: doubling it doubles the ending balance in this model. Starting ten years earlier does not universally halve the required monthly amount.

At a zero real return, the 42-year scenario ends at its 100,800 contributed; at an assumed 4% it ends near 256,146. Actual returns need not be positive or constant. Losses, taxes, fees, interruptions and withdrawal timing can change the result. These are not forecasts or evidence that 100,000 is enough to retire.

Use the compound-interest calculator for stated assumptions, checking its compounding and contribution conventions before comparing results. Nominal versus real values explains purchasing-power units; investment fees explains omitted costs.

Build a retirement page from documents

Collect public and workplace pension statements, contribution records and personal asset balances. Record projected income, commencement dates and assumptions for each source. Separate current rules from proposed reforms. Check employer contributions, vesting and access terms rather than assuming a benefit applies.

List essential and flexible retirement costs on a consistent price basis. Explore later access, lower returns, interrupted contributions and longer withdrawals. A later retirement date can change a model, but health and employment can limit the ability to work longer. Keep near-term obligations and an emergency reserve in view when assessing contributions.

Update the page when your household, scheme or assumptions change. The aim is a plan you can inspect and revise, with visible uncertainty rather than a promise of retirement or a prediction that it is impossible.