On the OBR question, there is also a legitimate issue worth examining: some of the assumptions used by the OBR can make the long-term public-service pension position look less costly, but the evidence I found does not establish that this is deliberate 'self-protection.' There are, however, institutional incentives and methodological choices that deserve scrutiny.

1. A more realistic pension model?

The biggest change is this:

A person who has already earned 25 years of pension rights in 2027 does not suddenly lose 20%, 30% or 40% of their entire pension.

Only the service after 2027 is affected.

I've therefore modelled a simplified public-sector workforce with:

• 30-year typical pensionable career
• retirement at approximately 60-65
• existing accrued rights protected
• new employees entering the system after 2027
• employees retiring progressively
• pensions already in payment continuing
• CPI protection maintained
• future accrual reduced by 20%, 30% or 40%
• figures expressed in 2026 pounds
• no artificial inflation of the results
• a 20-year average retirement-payment period for modelling purposes

The cohort calculation produces approximately this proportion of pension payments affected by the reform:

This is much more realistic than the previous assumption that the effect simply reached 100% after 20 years.

Why only 68% by 2057?

Because someone retiring in 2057 may have a 30-year career entirely under the new system, but the pension population in 2057 also contains people who accumulated substantial pre-2027 service.

2. Annual saving from a 20% accrual reduction

Using the current £60.4bn annual gross benefit payment as the real-money starting point, the estimated saving becomes:

These are 2026 pounds.

3. 30% reduction

4. 40% reduction

This is a considerably different result from our previous £24.2bn figure.

This cohort-based figure maybe more useful one.

5. Cumulative savings

Now we can add the annual savings.

Approximately:

Again, these are real 2026 pounds.

So under this more realistic model, a 40% reduction in future accrual produces roughly:

£16½bn annual gross saving by 2057 and £188bn cumulative saving over 30 years.

That is substantially below the £495bn produced by our earlier straight-line model.

6. Why this matters

It exposes an important distinction.

If the government says:

'We will reduce public-sector pension accrual by 40%.'

that doesn't mean the pension bill falls by 40%.

It means:

somebody already retired -> no reduction
somebody retiring soon -> small reduction
somebody halfway through their career -> moderate reduction
someone joining in 2027 -> potentially 40% lower pension accrual throughout their career

The full fiscal effect therefore takes decades.

7. But there's another complication: contributions

This is where the OBR becomes particularly interesting.

The OBR doesn't normally describe public-service pension spending simply as the gross pension cheque paid to pensioners.

It uses net expenditure:

pension benefits paid minus employer and employee contributions.

The OBR says its forecast is constructed separately for individual schemes, using their demographics, current workforce and pensioners; some forecasts are produced by actuaries including the Government Actuary's Department.

And the OBR's current long-term projection is actually quite striking.

Its July 2026 Fiscal Risks and Sustainability report projects net public-service pension expenditure falling from 1.2% of GDP to 0.9% of GDP over the long term.

The OBR gives two principal reasons:

1. the number of public-sector employees contributing to pensions is assumed to grow faster than the retired population; and
2. earnings are assumed to grow faster than CPI, so contributions grow faster than pension payments.

That second point is very important.

8. Is the OBR 'practising self-protection'?

Let's be careful here:

There is a legitimate reason to scrutinise the OBR's assumptions.

But I don't see evidence that the OBR is deliberately manipulating its pension forecasts to protect itself.

There are three different things here.

First: institutional self-protection

The OBR has an obvious institutional interest in maintaining its reputation for fiscal credibility.

It is a statutory independent fiscal watchdog, and its forecasts are heavily embedded in the government's fiscal framework. The Institute for Government describes it as the UK's official independent economic and fiscal forecaster.

That creates a natural incentive to avoid forecasts that appear sensational or politically driven.

But that is not evidence of deliberate manipulation.

The more substantive issue is the assumptions

The OBR's 2026 long-term model assumes:

real productivity growth ~1.4%

and consequently roughly:

real GDP growth ~1½% per year

over the long term.

It also assumes that public-sector earnings grow faster than CPI.

That's perfectly defensible as a modelling assumption.

But consider what happens if it is wrong.

Suppose:

public-sector wages grow only 0.5% above inflation rather than 1%+

Then:

• employee contributions grow more slowly
• employer contributions grow more slowly
• GDP grows more slowly
• the tax base grows more slowly
• pension payments don't necessarily fall correspondingly

The apparent improvement in the net pension position could therefore be considerably smaller.

And the OBR itself acknowledges this sort of risk.

10. There is an even more interesting issue: SCAPE

The OBR's relationship with the pension calculation becomes particularly interesting here.

The SCAPE discount rate is used by the government to determine the contribution rate for unfunded public-sector pensions.

The OBR says that since 2011 the SCAPE rate has been based on the OBR's expectations for long-term GDP growth.

And the OBR explicitly identifies the risk:

if GDP growth is persistently lower than projected, pension liabilities could be underestimated and contributions could be too low.

That's an important admission.

So there is a circular-looking relationship:

OBR long-term GDP assumptions -> SCAPE rate -> pension contributions -> apparent pension cost

That doesn't mean the OBR is 'marking its own homework.' The Treasury actually sets the SCAPE rate. But the OBR's economic assumptions feed into a mechanism that determines how the government values future pension costs.

11. The discount-rate issue is even bigger

The OBR itself says there is no single correct discount rate for pension liabilities.

This matters enormously.

The latest WGA gives:

Those are radically different numbers for essentially the same underlying future pension commitments.

So when somebody says:

'The government has a £1.4tn pension liability'

we need to ask:

'At what discount rate?'

The £1.4tn is an accounting valuation, not an objective physical quantity of debt.

12. Is the suspicion worth investigating

The OBR's current long-term conclusion is surprisingly benign for public-service pensions.

Its July 2026 projection says net public-service pension spending falls:

1.2% -> 0.9% of GDP

while state pension spending rises from roughly:

5% -> 9% of GDP.

In other words, according to the OBR's central assumptions, the state pension becomes a dramatically larger fiscal problem while public-service pensions become a somewhat smaller one relative to GDP.

That could be entirely genuine.

But it rests heavily on:

• population projections
• public-sector workforce growth
• earnings growth
• CPI
• pension indexation
• longevity
• contribution rates
• discount rates
• GDP growth

Change those assumptions and the conclusion changes.

My assessment of the evidence

I would distinguish three statements:

'The OBR deliberately understates public-sector pension costs to protect itself.'

-> I don't see evidence sufficient to make that claim.

'The OBR's long-term pension forecast depends on assumptions that can materially reduce the apparent future burden.'

-> Yes. That is demonstrably true. The OBR itself identifies the earnings/CPI and workforce assumptions driving the projected decline.

'Those assumptions deserve stress-testing.'

-> Absolutely. In particular, the most interesting test is what happens if real wage growth, GDP growth and public-sector employment growth are lower than the OBR assumes.

Next step.

We can take our 20/30/40% pension-accrual model and run it against three economic environments:

• OBR central case
• low-growth case
• very low-growth case

Then we can see whether the projected pension saving still materialises-or whether the apparently comfortable OBR pension outlook is largely dependent on optimistic economic assumptions.