Jon Spain – author and former erudite GAD actuary.
Revisited : Approximate Costs Over 15 Years From 2024/25
Outline
- understanding large numbers
- triple lock background
- major UK state expenditure 2024/25
- UK state pension
- pensioner population
- triple lock definition
- wages growth
- prices growth
- random numbers used
- results
- conclusions
Understanding Large Numbers One of my former bosses used to tell me that senior politicians often found very large numbers hard to comprehend. His solution was to talk about a local hospital building or a fraction of the Channel Tunnel. My unit will be Education (£111 b). On the other hand, someone is apparently prepared to pay $8 m for a World Cup final ticket!
Triple Lock Background Introduced by the coalition government, the state pension triple lock (see definition below) was announced in the 2010 Budget and was officially implemented for the 2011/12 financial year. While the cost has been described as prohibitive, politicians have been reluctant to upset voters. During 2024/25, the total UK state pension amounted to £138b, which is huge. How much does the triple lock actually cost?
Major UK State Expenditure 2024/25 The amount is £1.17 trillion. One HMRC breakdown has figures adding up to £1.01 trillion, with the top 5 line items (in £ billions) being 297 (Work and Pensions). 216 (Health and Social Care), 111 (Education), 57 (Defence) and 50 (Scottish Government). Taken together, these 5 items represent 72% of total government expenditure.
UK State Pension During 2024/25, the amount paid was £138 b, which was 46.4% of the Work and Pensions expenditure and 13.6% of total expenditure. This cost is major. It came to around 5% of GDP, 35% higher than 50 years earlier and 15% higher than in 2010/11. In around 50 years time (early 2070s), in their Fiscal Risks & Sustainability Report (July 2025), OBR estimated that the state pension cost would increase to 7.7% of GDP. This is unsustainable.
Pensioner Population The state pension was substantially upgraded for new awards from 2016. With more pensioner awards being made under the new system, just starting from the 2024/25 position tends to understate the costs. On the other hand, I have excluded mortality, which tends to overstate the cost. Without data, I have simply assumed a stationary pensioner population, with new awards being balanced by deaths. The OBR estimates (see above) are way more accurate than mine but I’m just trying to assess an order of magnitude.
Triple Lock Definition The state pension is increased by the largest of 2½%, wages or inflation; it is never reduced. Respectively, the latter two metrics are defined as the average increase in weekly wages across the UK (measured May to July of the previous year) and the rate of CPI inflation in the 12 months leading up to the previous September.
Wages Growth For my estimates, I have looked at wages over calendar years. When I have time, I shall look at how far that diverges from the actual definition used but I doubt that there will have been a significant difference. Between end-1974 and end-2025, wages increased on average by 6.2% pa over 1 year and by 5.1% pa over periods of 15 years (see below).
Prices Growth For my estimates, I have looked at CPI over calendar years. Some years ago, I looked at RPI seasonality, which was insignificant; I don’t believe that CPI will show anything different. Between end-1974 and end-2025, CPI increased on average by 4.2% pa over 1 year and by 3.2% pa over periods of 15 years (see below).
Random Numbers Used Deriving 15,000 correlated values leads to having 1,000 values for each of 15 years, calibrated (means and standard deviations) to the original data. The distributions used were Extreme Value for wages and Inverse Gauss for prices.
Results Depending upon which increases are assumed to be awarded, I have projected the payments over 15 years. The two measures I have extracted are the total expenditure and the ratio of total expenditure to the base case (zero increases). For each measure, I have calculated the mean, the standard deviation, the minimum and the maximum. There is an interactive dashboard showing all statistics at the URL below but below I show a table based upon means.
https://www.jonspain.com/TripleLockJon_Jul2026/
Triple Lock Additional Cost The cost impact of 2½% fixed increases (405) is slightly over one-half that of prices alone (734) and one-third of wages alone (1,210). When combined with either prices or wages, the addition of a fixed increase has little impact. Over the next 15 years, against assuming no increases at all, the cost would be £1.4 t but that seems unrealistic. In reality, one can expect the state pension to be increased every year, say in line with CPI or CPIH. While not identical in each year, they tend to converge over longer periods.
Conclusions Were the triple lock retained for 15 years, then the average total outgo would be £3.5 t, the final year’s average payment being £357 b, which is 2.6 times the original £138 b. It seems somewhat unlikely that the GDP would have grown as fast. Realistically, any solution might include CPI alone, in which case the average total outgo would still be £2.8 t, the final year’s average payment being £245 b, which is 1.8 times the original £138 b. This is comparable to the increase from 5% to 7.8% (of GDP) in the 2070s (see above).
I have every respect for Jon Spain and his work. The figures here are interesting and I am totally in favour of open debate about the future of the State Pension and how it will increase in future. What I do not understand is why he says in the middle of the piece “This is unsustainable”. This simply a statement of opinion with no supporting arguments. What is sustainable in 15 years time is a matter for politics, not mathematics.
A zero-uprating baseline is purely hypothetical and policy‑irrelevant. No serious fiscal analysis would treat freezing the nominal state pension for 15 years as an appropriate counterfactual. A much more meaningful comparison is between triple lock and some credible alternative, typically earnings or price indexation.
The long‑run annual averages (wages 6.2% one‑year, 5.1% over 15‑year windows; CPI 4.2% and 3.2%) are drawn from a rear view mirror including the high‑inflation 1970s, the early 1980s disinflation and are largely the result of pre‑inflation‑target regimes. The more recent macro regime (an “independent” central bank, explicit inflation targeting, post‑2008 levels of “neutral rate of interest r*” for major economies like the US, UK, and Eurozone which have dropped significantly and stayed remarkably low) suggests future average CPI closer to 2–3% and nominal earnings closer perhaps to 3.5–4.5% in projections from OBR/BoE of recent years, not a mechanical repeat of 50‑year history. Using a backward-looking 50‑year distribution as if it were stationary is at best only a very crude approximation.
OECD and G7 comparisons show the UK currently spends near the bottom of the league on public pension benefits as a share of GDP (around 4.7–4.8% in recent data, versus eg Italy around 12–13% and France around 14% on wider old‑age benefits). Even if UK spending rose to 7–8% of GDP over 50 years, that would still leave us below or around many current continental levels, not in uncharted territory.
UK state pensions, taken on their own, provide one of the lowest replacement rates among advanced economies, even though overall many old‑age incomes in the UK may be boosted at least to some extent by private pensions.
Jon Spain’s piece is essentially a stylised stress‑test demonstrating “large numbers” rather than a serious fiscal‑sustainability analysis, and it quietly bakes in a very harsh and unrealistic baseline (no uprating) to make the triple lock look extravagant. It also imports an “unsustainable” verdict from OBR’s long‑term projections without really engaging with some of the levers (changes to SPA, contributions, tax base) and known unknowns (for one example, AI may in time be of a magnitude of something we have not seen in our lifetimes, in terms of its potential transformational effects on the global economy) that determine sustainability in practice.