The Price They Pay Themselves

There is a number the British state has decided is the right amount to put aside for a decent retirement. It is 28.97% of salary. We know this because the state pays it, every month, into the pensions of the people who work for it.

There is a second number, which the same state has decided is the right amount to require of everyone else. It is 3%. And it is not even 3% of salary — it is 3% of the slice of pay between £6,240 and £50,270, which for a worker on £60,000 comes to a shade over 2% of what they actually earn.

Both numbers were set by the same government, with the same actuarial advice, on the same evidence, in the same decade. Nobody has ever been asked to explain why they are different by a factor of thirteen.

What thirteen times buys

Take a worker who starts at £18,000 at twenty-five and finishes on £60,000 at sixty-six. Forty-two years, no career breaks, one salary.

Put her in the Civil Service alpha scheme and she retires on £49,348 a year, index-linked, guaranteed, with a survivor's pension attached. Add the State Pension and she has £61,896 — one hundred and three per cent of her final salary. She will be better off in retirement than she was in work.

Put the same woman in a private job paying the statutory minimum and she retires on £11,153 from her pension pot. Add the State Pension and she has £23,701. Thirty-nine per cent of final salary.

Same career, same salary, same forty-two years. One of them retires on more than she earned; the other on less than half.

Now the part that should end the argument. The civil servant contributes 5.4% of her own pay across that career. The private worker contributes 3.6%. The woman who ends up with the smaller pension is not the one who saved less of her own money. What she lacked was an employer prepared to put in 28.97% — an employer she was funding out of her taxes regardless.

The number they will not say out loud

The state has published its view of what an adequate retirement looks like. The Second Pensions Commission's target replacement rate for a £60,000 earner is 54% of pre-retirement income. That is the official standard, the yardstick by which the Department for Work and Pensions has concluded that fifteen million working-age people in this country are not saving enough.

Measure both our workers against it. The private employee lands at 39%, fifteen points below the government's own adequacy line. The civil servant lands at 103%, forty-nine points above it.

Read from the contribution side, the same fact: buying the civil servant's outcome privately would take roughly 25% of full salary for forty-two years. Hitting the government's stated adequacy target would take about 10%. The government mandates the equivalent of 5.8%.

They know what adequacy costs. They have modelled it, published it, and built an entire Commission around the fact that the country is not achieving it. Having established the number, they mandate barely half of it for the private sector and pay two and a half times it to themselves.

The excuses, and why they fail

"The 28.97% isn't a choice — it's an actuarial valuation." It is the output of a valuation of a promise the state designed. The 2015 reforms were a deliberate act, taken by a Treasury that knew precisely what each accrual rate would cost, in the same period that the same Treasury was setting the auto-enrolment phasing schedule. Nor is the choice historic. It is remade at every valuation, including the one running now, which will set rates from April 2027. Every four years the state examines the promise and keeps it.

"Pension is deferred pay. It's an employment matter, not a citizenship matter." It would be, if the money came from shareholders. When Aviva pays 12%, that is Aviva's owners choosing what to do with Aviva's profits. When the state pays 28.97%, that is the woman on 2% of her salary paying for it. She is not a shareholder. She was not consulted. She cannot decline the relationship.

"We can't raise the minimum — employers can't afford it." This is the excuse that gives the game away, because it is true and it is precisely the point. The Institute for Fiscal Studies has modelled the options: raising minimum contributions to 12% from the first pound would generate £17 billion a year in additional saving, and cut low earners' take-home pay by 4% once employers pass the cost through into wages. That is a serious constraint on a real hospitality firm with real margins.

It is not a constraint that applies to the state, which faces no solvency test, no margin, no moment where the promise breaks the balance sheet. It writes an unfunded promise, values it on a discount rate it sets itself, and sends the bill forward. Affordability is a discipline the government imposes on everyone except itself, and then cites as the reason it cannot help.

The asymmetry, stated plainly

For a private sector worker, a better pension means a smaller wage packet this month. That trade-off is visible, immediate and politically painful.

For a public sector worker, a better pension means a higher tax bill for somebody else, some time later. That trade-off is invisible, deferred and politically free.

Two identical objectives, two completely different distributions of who pays — and the group bearing the visible cost is also the group funding the invisible one.

That is the bias. Not a conspiracy, something duller and worse: an institution that has, over three decades, made every choice about its own employees on the assumption that cost is a problem for the future, and every choice about everybody else's on the assumption that cost is a problem for right now. Nobody in Whitehall owns the seam between the two. The valuations sit with the Treasury and the Government Actuary's Department. Auto-enrolment sits with the DWP. No select committee has ever put the numbers on the same page.

What would actually answer it

Not levelling down: cutting alpha releases no cash quickly, because accrued rights are protected — McCloud settled that — and unfunded schemes turn reduced accrual into a smaller future liability rather than money today. Not levelling up to 28.97% either; nobody has proposed it, because employers who face actual constraints cannot bear it.

What can be answered is the question itself. If 54% is the standard, explain why one group of citizens is funded to nearly twice it and another to well under it. If the answer is that the first group are employees and the second are merely the public, say that out loud and let the public hear it.

Then start with the easy part. The legislation removing the lower earnings limit received Royal Assent in 2023. Commencing it would lift that private worker's pension by 42% at a stroke. Three years on, the regulations have not been laid, and the Pensions Minister has confirmed there will be no auto-enrolment changes this Parliament.

A government that cannot find the will to commence its own three-year-old Act, while paying 28.97% into its own staff's pensions without pause or debate, has told you where its priorities sit. It did not need to say a word.

Figures. Civil Service employer contribution 28.97% of pensionable pay from 1 April 2024 (Cabinet Office, Civil Superannuation). NHS 23.7%. Auto-enrolment minimums 3% employer / 8% total of qualifying earnings £6,240–£50,270. Projections assume 2.98% nominal salary growth, CPI 2%, 5% net DC investment return and a 5% index-linked annuity rate at 67; alpha accrual 2.32% revalued at CPI. Target replacement rates per the Second Pensions Commission interim report. Reform costings and wage pass-through estimates from IFS, Automatic enrolment: trends in employer pension contributions and the impact of potential reforms, July 2026.