Inflation after the salary stops

What changes on the day the salary stops

Inflation bites harder after retirement because a salary reprices and a fixed income does not. While you work, a rise in prices is met — imperfectly, and with a lag — by a rise in pay. Once you stop, the same rise lands entirely on your standard of living and stays there.

That is the whole of it, and it is worth separating from the usual framing. The common version says prices rise faster for retirees. That is a separate argument, it is dealt with further down, and it is not the main one. The main one is that the offset disappears, not that the rate changes.

Think about what you actually own while you are earning. There is the portfolio, and there is a claim on your own future work. The second one adjusts. Not annually, not reliably, not for everyone — but across a working life, pay tends to move with the price level in a way that a declared deposit rate does not.

Retirement is the day you dispose of that second asset. What is left is a set of claims denominated in rupees, held by someone whose spending is denominated in goods. Those two things drift apart every year, and after you stop working nothing pulls them back together.

None of this is an argument about how much inflation there will be. It is an argument about what a given amount of it does to two people in different positions. What inflation is, and how a real return is worked out from a nominal one, belongs to the inflation explainer; this article assumes it.

The four kinds of claim, and which one you just gave up

Sort everything a household can hold by one question: when the price level moves, does the rupee value of this claim move with it? The answer splits into four groups, and the last row is the one that changes at retirement.

Read the table downward and the position of a working person looks different from the way it is usually described. The largest inflation-linked asset most people own is not in the portfolio at all. It is their own earning power, and it is doing quiet work that nobody thanks it for.

Retirement removes that row. Whatever is left has to carry the whole job on its own — which is why the question of what the remaining rows are made of stops being a matter of taste and becomes the structural question of the retirement. That structure is the subject of layering retirement income.

What thirty years does to a fixed number

Take 6% a year. It is an arbitrary round number, chosen because it divides neatly and because you can redo every line below on paper — it is not a forecast, and a different rate gives a different answer. What does not change with the rate is the shape.

Start with the doubling shortcut. Divide 72 by the rate and you get the years a price level takes to double — 72 divided by 6 is 12. Prices doubling and a fixed income halving in what it buys are one event described twice, so the same twelve years does both. That single line is the rule of 72 doing the entire argument.

Now run it out. At 6%, prices multiply by roughly 5.74 over thirty years, because 1.06 multiplied by itself thirty times comes to 5.74. So spending of ₹60,000 a month today needs about ₹3.44 lakh a month in year thirty to buy the same things.

Turn that around and it is the sentence a retiree actually needs. A pension fixed at ₹60,000 a month buys, in year thirty, what about ₹10,450 buys today. It passes through ₹29,800 at year twelve and ₹14,800 at year twenty-four on the way — half, then a quarter, at a rate nobody would call dramatic in any single year.

Two things about that progression are worth pausing on. It is smooth, so no year is ever the year it becomes a problem. And it is arithmetic rather than misfortune — nothing has to go wrong for it to happen, which is what separates it from every other risk on a retiree's list.

The corpus side of the same calculation — how large a starting sum has to be to fund an expense that grows this way — is the corpus question, and it is a different article.

The basket reweights while the prices move

Now the second effect, the one the opening set aside. Two things happen at once as a retirement goes on, and they multiply rather than add.

The first is ordinary price rise, applied to everything. The second is that the basket itself changes shape: in the ordinary case the share of a household's spending that goes on health, medicines, tests, help in the home and eventually care rises with age, while the share going on commuting, work clothes, education and supporting children falls away.

That reweighting matters because a published index is built on an average household's basket, and an eighty-year-old's basket is not an average household's. Your own rate of price rise is a weighted average of the categories you actually buy — so a retiree whose spending has tilted toward the categories that rise fastest experiences a higher personal rate than the headline, without the headline being wrong about anything.

This is the point at which most writing reaches for a medical-inflation figure. There is no verified one here, so there is none in this article. The mechanism does not need one: two effects compounding is worse than one, whatever the sizes.

Health cover is where the effect becomes visible in a bill you can see, and there is one hard number that is worth knowing. For indemnity individual health cover held by someone aged 60 and above, the premium revision is capped at 10% per annum. Read that as a ceiling and not as a forecast — but note what a ceiling of that size permits: run the doubling shortcut again and a premium rising at the cap doubles in a little over seven years. The ceiling, not the expectation, and it is still a doubling inside a decade.

The consequence for planning is unglamorous. A retirement budget built on today's premium and today's medical spending is a budget built on the lowest figure it will ever contain, and the categories doing the rising are the ones that are hardest to cut. What health cover is for, and what it does not cover, is the health insurance guide.

Why an all-deposit portfolio is not the safe choice it looks like

Because every rate in it is a declared nominal rate, and not one of them reprices when the price level does. Set the currently declared rates side by side, since the argument is easier to make with real numbers in front of it than in the abstract.

• The Public Provident Fund pays 7.1%.
• The Senior Citizens' Savings Scheme pays 8.2%.
• The Post Office Monthly Income Scheme pays 7.4%; its five-year recurring deposit 6.7%; its time deposits 6.9% for one year, 7.0% for two, 7.1% for three and 7.5% for five; and a Post Office savings account 4%.
• The National Savings Certificate pays 7.7%, and Kisan Vikas Patra 7.5%, maturing in 115 months.
• The Employees' Provident Fund credits 8.25% for FY 2025-26 — the financial year to March 2026.

That last one deserves a sentence of its own, because it is reported carelessly more often than not. The EPFO Central Board of Trustees recommends a rate; the Central Government ratifies it under para 60(1) of the EPF Scheme 1952; EPFO then directs the credit. Recommended, ratified, then credited — three stages, and a headline reporting the first as though it were the third is reporting something that has not happened yet. The figure quoted above has been through all three.

Now the point of listing them. Every one of those figures is a number of rupees per hundred rupees per year. Not one of them is expressed as a margin over a price index. They are declared for a period and re-announced for the next one, and a rate that has to be re-announced is not a rate that tracks anything automatically. Nominal, not indexed.

Here is the specific mistake, and it is the commonest one on this topic. A saver compares a declared rate with a headline inflation reading, finds the first above the second, and concludes the money is growing in real terms. Two things break that comparison. Whatever tax applies to the interest comes out first, so the rate being compared is not the rate being received. And the index is the average household's basket, as the previous section argued, rather than the basket of the person doing the comparing. The gap that looked comfortable can be much thinner than it appeared, or the other way round.

So the deposit-only portfolio has not removed risk from a thirty-year retirement. It has chosen which risk to run — and it has chosen the one that is invisible on a statement, arrives without a bad day, and cannot recover. One risk swapped for another, and the swap is worth making deliberately rather than by default.

Why working longer only half-answers this one

Because it shortens the exposure and leaves the rate alone. Fewer years of retirement means fewer years of erosion, and that is arithmetic rather than a rhetorical concession. What more working years cannot do is change the rate at which the years that remain erode. Shortens exposure, not erosion.

That makes inflation unusual among retirement problems, because most of them do yield to the same lever. A corpus that is short can be topped up. A late start can be partly caught up. A bad market in the year before you planned to stop can be waited out. In each of those, more working years is a whole fix; here it is half of one.

And the exposure it shortens is bounded by the one input you do not control. How much you save is a decision. How long you live after you stop is not. A plan built to an average life expectancy is, by construction, one that a large share of that cohort outlives — and the years past the average are precisely the years in which the price level has compounded the longest.

There is a second-order effect that cuts the other way and should be admitted. Retiring later also means the retirement starts at a higher price level, so the first year's expense is larger in rupees even though it buys the same things. Working longer therefore moves the starting point as well as the finish line. It is still a real fix for the size of the corpus. It is simply not a fix for the mechanism.

Which leaves the uncomfortable conclusion this article exists to state plainly: of the risks a retirement runs, this is the one where effort and discipline before retirement buy the least protection. The protection has to be built into what the money is held in, not into how long it was earned for.

What the alternative costs, stated in full

It costs a fall that can land in the year the money is needed. The property that lets an asset keep pace with prices is that its cash flows are not fixed — and that is the same property that lets it fall hard. Selling into a fall means selling more units to raise the same rupees, and those units are not there for the recovery — which is the mechanism set out in the withdrawal-rate article and it is entirely real. Nothing here is free.

So none of this says deposits are wrong and equity is right; that would be the same error with the sign flipped. The retiree is choosing between two failure modes, not between risk and safety. One fails slowly, invisibly, and without any possibility of reversal. The other fails suddenly, visibly, and can reverse — but only if you are not forced to sell while it is down. Stated that way, the design problem becomes obvious: hold enough of the second kind to matter over thirty years, and arrange things so that you are never forced to sell it at the wrong moment.

That second half is a structural question with a standard answer, and it is not this article's to answer. Keeping the next few years of spending in something that cannot fall, so the long-horizon assets are never sold under pressure, is the bucket strategy. How the withdrawals themselves behave is the systematic withdrawal article. What the overall split between the kinds of claim should look like is asset allocation.

What this article claims is narrower and, we think, harder to argue with: a retirement measured in decades has to hold something whose value is not fixed in rupees, because everything that is fixed in rupees is losing ground by construction. How much, and held how, is a question about a specific person's income, health, dependants and tolerance — which an article does not have.

Checking the size of the effect rather than arguing about it

The mechanism above is arithmetic and settles itself. How large the effect has actually been, and how much the answer moves depending on when a retirement started, is a question you put to a price history instead of to an opinion.

FNOTrader's Mutual Funds app runs on the full AMFI history of per-unit prices — the NAV — around 34 million rows of it, and reports rolling-return distributions across every available start date alongside maximum drawdown and how long recoveries took. Those are the two figures a retiree's version of this question turns on: how far the value fell, and how long it stayed down.

The step it does not do is the one this article is about. Every figure it reports is a nominal one, and the subtraction of an assumed rate of price rise from it has to happen before any of those figures mean anything — because a return that has not had inflation taken out of it is not a return anyone can spend. Historical outcomes describe what happened rather than what will happen, and past performance does not indicate future results.

FNOTrader is not a SEBI-registered investment adviser and none of this is retirement advice. What a particular person should hold depends on facts an article does not have.

Common questions

Continue reading

More in Retirement Planning · App: Mutual Funds · Definitions: glossary · Free tools: calculators · All: every article