The Triple Lock

Prepare for a long read. I heard someone ask the other day, quite reasonably, why pensioners get a Triple Lock while the working population does not.

And on the face of it, it is a fair question. If pensions rise by whichever is highest - inflation, wages or 2.5% - why should wages not get the same treatment? Why should the retired get the nice little ratchet mechanism while the working population gets a performance review, a motivational poster in the staff kitchen, and a reminder that “times are difficult”?

It sounds fair. It sounds symmetrical. It has that attractive pub-logic quality where everyone nods for about thirty seconds before someone remembers the electricity bill.

The first problem is that wages are not just income. Wages are also costs. If every worker’s pay is automatically lifted by the highest of inflation, earnings growth or 2.5%, then every employer’s wage bill rises whether the business can afford it or not. The big firms grumble and pass some of it on. The marginal ones cut hours, delay hiring, raise prices, or simply disappear from the high street, to be replaced by a vape shop and a sense of decline.

And then, because prices have risen, inflation rises. And because inflation has risen, wages rise again. At which point the whole thing starts to look less like fairness and more like a circular argument with direct debit facilities.

Pensions are different in one important respect. They do not directly increase the cost of producing a loaf of bread, running a care home, staffing a hotel, or keeping a restaurant open on a wet Tuesday in February. But they absolutely do increase spending power. Pensioners buy food, energy, insurance, holidays, meals out, garden supplies, Christmas presents and occasionally suspiciously expensive bird seed. So of course pension uprating contributes to demand. It is not inflation-free fairy dust sprinkled from the cardigan cupboard.

The awkward bit is that the State Pension is not a tiny corner of the welfare state. It is the biggest bit. DWP figures say around 55% of social security spending goes to pensioners. In 2025/26, pensioner benefits are forecast at £177.7 billion, including £146.1 billion on the State Pension. There were 13.2 million people receiving the State Pension in August 2025.

So when politicians talk about “benefits”, they often do so in a tone that suggests a man in a tracksuit smoking outside a betting shop. But the largest slice of the benefits bill goes to pensioners. That is not an opinion. It is arithmetic, which is always inconvenient when everyone has turned up carrying a prejudice.

So we have ended up with a strange moral hierarchy. A pension rising automatically is “dignity in retirement”. A working-age benefit rising properly is “dependency”. A public sector worker asking not to get poorer is “unaffordable”. A young renter facing half their income disappearing into someone else’s buy-to-let is told to stop buying coffee, which would at least be more convincing if a flat now cost the same multiple of earnings as it did when the advice-giver bought one.

The Triple Lock may have made sense when pensioner poverty was the great neglected problem. It was crude, but politics often likes crude tools because they fit neatly on a leaflet. The trouble is that once a crude policy becomes a sacred object, nobody is allowed to ask whether it is still doing the same job.

Because pensioners are not all poor. Some are, and they need proper protection. But some own homes outright, have occupational pensions, savings, investments, and a winter fuel payment debate conducted with the moral delicacy of a man defending his second helping of pudding. Meanwhile, plenty of working people are paying rent, childcare, commuting costs, council tax, energy bills, student debt, and National Insurance, while being told that asking for security is fiscally reckless.

There is also the small matter that pensioners vote. Reliably. In large numbers. With the grim punctuality of people who still know where their polling card is and regard civic duty as something slightly more important than updating TikTok. They are also a large and growing voter bloc, which means the Triple Lock is not merely a pensions policy. It is an electoral forcefield. Touch it and every party strategist starts hearing distant church bells and the sound of marginal seats falling over.

But underneath all that lies the bigger problem, which is not sentiment, fairness or even electoral cowardice. It is demography.

The workforce is not growing fast enough to carry an ever-growing retired population on ever-more generous terms. The State Pension is not a savings account with everyone’s contributions sitting in a labelled biscuit tin somewhere in Whitehall. Today’s pensions are largely paid from today’s taxes. That means today’s workers are paying for today’s pensioners, just as tomorrow’s workers will be expected to pay for us.

That arrangement works tolerably well when there are plenty of workers for every pensioner. It becomes rather more awkward when the number of pensioners rises and the working-age population does not keep pace. At that point the promise does not vanish. It just moves. It lands on younger workers through tax, on public services through cuts, on the national debt through borrowing, or on future pensioners through later retirement and lower uprating.

This is the bit politicians dislike saying out loud. The Triple Lock is not just a kindness to current pensioners. It is a claim on the future earnings of a workforce that is smaller than we would like, less securely housed than it used to be, and already carrying tuition debt, childcare costs, rent, National Insurance, frozen tax thresholds and the general pleasure of being told to work harder by people who bought houses when they cost about three decent salaries and a packet of crisps.

You can raise the pension age, but that hits manual workers and the less healthy hardest. You can increase immigration, but then half the same politicians who defend the Triple Lock start shouting about numbers. You can raise productivity, which is the elegant answer, except Britain has been promising that since roughly the invention of the fax machine. Or you can trim the generosity of pension uprating for those who do not need it. That is not cruelty. That is arithmetic arriving with its coat on.

So the question “why not give workers a Triple Lock too?” is useful, not because the answer is “yes, let’s do that”, but because it exposes the oddness of the original lock. If applying the same principle to wages would be inflationary, unaffordable and economically dangerous, perhaps we should at least admit that applying it permanently to the largest benefit in the system is not an act of divine wisdom either.

And if one of the three locks has to go, it should be the 2.5% floor.

The earnings link has a logic. It stops pensioners falling behind the working population over time. The inflation link also has a logic. It stops pensioners being made poorer in real terms when prices rise. But the 2.5% guarantee is the odd one. It says pensions should rise by at least 2.5% even when inflation is low and wages are barely moving. That is not protection against poverty. That is a ratchet.

A Double Lock of earnings or inflation would still protect pensioners from falling behind either prices or workers. It would remove the extra upward tick that keeps lifting the baseline even in years when the economy has not really justified it. The Institute for Fiscal Studies has estimated that the Triple Lock could add somewhere between £5 billion and £40 billion a year to State Pension spending by 2050, compared with earnings indexation, depending on how the economy performs. That is quite a large uncertainty range, admittedly, but even the bottom end is not exactly loose change found in the Treasury sofa.

Then comes the nastier question: should the State Pension itself be means-tested or tapered?

The politically survivable answer is to start any taper at about £50,000 of taxable income. That sounds neat because it roughly tracks the higher-rate tax threshold. It also avoids hitting people on quite ordinary retirement incomes. But it would not save enough to matter very much. It would be reform by press release.

The necessary answer is lower, and therefore much uglier. If the country actually wanted serious savings, the taper would probably have to start somewhere around £30,000 to £35,000 of other taxable income, excluding the State Pension itself. Then it could be withdrawn at, say, £1 of State Pension for every £3 of income above that. Since the full new State Pension is now £241.30 a week, about £12,548 a year, that would mean someone starting the taper at £30,000 would lose the full amount once their other income reached roughly £68,000. Start at £35,000 and the full loss comes at about £73,000.

That is the zone where the savings become real. Not theoretical. Not symbolic. Real.

The exact saving would need proper DWP and HMRC modelling, because it depends on whether you assess individuals or households, how you treat savings, how many people rearrange their income, and how many lawyers suddenly develop a touching interest in pension planning. These are not official savings estimates, but they show the scale. If you withdrew an average of £2,500 from two million better-off pensioners, that is £5 billion a year. If you withdrew an average of £5,000 from two million, that is £10 billion. If four million pensioners lost an average of £4,000, that is £16 billion. That is enough to matter.

That is the hard version. It saves real money, but it also hits people who did what governments told them to do: save, contribute, build a private pension, and do not rely entirely on the state. It would not only hit the very rich. It would hit retired teachers, police officers, civil servants, managers, engineers, small landlords, widows with inherited pensions, and a fair number of people who thought prudence was supposed to be rewarded rather than entered on a spreadsheet as a recoverable offence.

There is a cleaner halfway house. Keep the Triple Lock for pensioners below, say, £50,000 of taxable income, but above that level reduce it to a single lock. Personally, I would make that single lock inflation. If you are living mainly on the State Pension, you need protection against inflation, wage growth and the general drift of living standards. If you are already on more than £50,000 a year in retirement, the state does not need to keep handing you the full ratchet. It only needs to stop the State Pension being eroded by prices.

That is not taking away the State Pension. It is not even freezing it. It is simply saying that the full Triple Lock is poverty protection, not a loyalty bonus for people with a good final salary pension and a conservatory.

To avoid a cliff edge, you could phase it. Full Triple Lock below £50,000, a blended uplift between £50,000 and £60,000, and inflation-only above £60,000. That way someone does not suddenly lose out because they crossed the line by £200 after cashing in a modest investment or receiving a slightly better occupational pension increase.

Would that be easy to apply? Not especially. It is easier than means-testing the whole State Pension, but not easy in the cheerful ministerial sense of “we have made a colourful flow chart”.

The main practical problem is that the State Pension is paid by DWP, while the relevant income information sits with HMRC. So you would need a routine annual data match between HMRC taxable income and DWP pension uprating. Technically possible, yes. Administratively elegant, no. This is Britain. We can make a paperclip require a login, a reference number and a forty-minute call queue.

The next problem is timing. Pension uprating is set before the tax year’s income is fully known. So the system would probably have to use the previous completed tax year’s income. That creates odd cases. Someone retires and has one unusually high income year. Someone sells investments and briefly crosses the line. Someone’s rental income varies. Someone draws down a pension pot one year and not the next. Someone’s income creeps above the threshold because their occupational pension rises with inflation.

So you would need smoothing, averaging, or an appeals process. Otherwise you would get the usual British administrative genius: a policy designed to save money, followed by a helpline playing Vivaldi to furious retired accountants.

There is also the question of whether the test is individual or household-based. The State Pension is an individual entitlement, but real household finances are often shared. A pensioner on £20,000 with a spouse on £90,000 is in a different position from a pensioner living alone on £20,000. If you assess households, the system becomes more intrusive. If you assess individuals, it is simpler but easier to game. There is no completely clean answer. There rarely is, which is why slogans are so much more popular than policy.

AI could make some of that easier. It could match DWP and HMRC records, spot unusual one-off income spikes, flag cases for human review, calculate which uprating rule applies, pre-fill forms, detect avoidance patterns, and explain decisions in plain English. This is precisely the sort of dull administrative plumbing where AI ought to be useful, rather than pretending to write poetry about brand values for a yoghurt company.

A decent system could say: this pensioner was below £50,000 last year, so the full Triple Lock applies. This one was between £50,000 and £60,000, so the blended uplift applies. This one was above £60,000, so inflation-only applies. This one made a one-off pension withdrawal or sold an asset, so a human being should look at it before the machine does something stupid in twelve-point font.

But AI does not answer the hard questions. It cannot decide the moral threshold. It cannot decide whether the test should be individual or household-based. It cannot decide whether a widow with inherited pension income is in the same position as a retired executive with rental properties. Those are political and ethical choices, not software bugs.

Nor should anyone want an algorithm quietly reducing pension upratings without proper audit trails, appeal rights and human override. Otherwise we simply replace “computer says no” with “algorithm says no”, which is the same misery with better marketing and a larger invoice.

Still, the principle is sound enough. Below £50,000, keep the Triple Lock. Between £50,000 and £60,000, phase it down. Above £60,000, inflation-only. Use taxable income, ignore capital itself, count taxable income from pensions, rent, savings and investments, and do it annually rather than turning it into a monthly bureaucratic pantomime.

And Britain should not pretend this is administratively unimaginable. Australia already does something much more direct. Its Age Pension is openly means-tested, using both an income test and an assets test. The family home is treated separately, but other income and assets are assessed, and the lower entitlement wins. It is not perfect, and I dare say it keeps a healthy number of accountants in shoes, but the sky has not fallen in. Australia still has pensioners, elections, suburbs, barbecues and people shouting at the television.

This matters because Britain has a habit of treating any serious discussion of pension means-testing as if someone has proposed putting a parking meter outside the Cenotaph. But other countries do make distinctions between poorer and richer pensioners. They do not all simply hand the same increase to everyone and call it fairness because it avoids a difficult conversation.

If ministers had courage, they would do the halfway house. Keep the State Pension universal, scrap the 2.5% Triple Lock floor, and limit the full uprating protection to those who actually need it. If ministers had only cunning, they would claw back more from affluent pensioners through the tax system and hope nobody noticed the plumbing.

The halfway-house option would not save as much as a hard means-test. But it would save something, and more importantly it would establish the principle that pension protection should be strongest where pension need is greatest. Which sounds dangerously like fairness, so naturally it would have to be introduced very quietly, possibly while everyone was distracted by a reshuffle, a royal balcony, or a minister getting trapped in a spreadsheet.

But none of this escapes the central problem. A growing pensioner population cannot be permanently insulated from economic reality while the bill is handed to a working population that is proportionately smaller, poorer in assets, and already being asked to fund everything else.

The real scandal is not that pensioners are protected. Poor pensioners should be protected. The scandal is pretending that age alone is a claim to ever-rising protection, while need, wealth, housing security and the number of workers paying the bill are treated as awkward details to be hidden behind a Union Jack and a campaign leaflet.

The trouble is that all these options require politics to distinguish between poor pensioners and comfortable pensioners, between need and entitlement, between fairness and bribery, between demographics and wishful thinking, and between economics and doorstep slogans.

So naturally we will not do that.

We will carry on calling pension spending “earned”, working-age support “welfare”, and wage demands “inflationary”, while a large automatic ratchet sits in the middle of the public finances wearing slippers and daring any politician to touch it.