Money in your forties

The decade the number becomes countable

In your forties the retirement calculation stops being a projection and becomes a count. You can observe your own spending, name a year you would like to stop earning, and multiply out how many contributions remain. The arithmetic does not change with age. The room to be wrong does.

Someone at forty-four planning to stop at sixty has 192 monthly instalments left. At thirty they had 360. That is the whole difference in one line, and it is worth writing down because it converts a vague sense of lateness into a number, not a feeling — which is the only form in which it can be acted on.

The second change is that the inputs are finally knowable. At twenty-eight, an estimate of retirement spending is a guess about a household that has not formed yet, a career that has not turned yet and a city you may not stay in. By the mid-forties most of that has resolved: you have several years of observable spending, and you know roughly which obligations end and which do not. The construction of the corpus figure itself is unchanged and is set out in the corpus arithmetic. What is different is that the answer is now built on measurements rather than on assumptions, which is what makes it worth doing carefully at forty-four and only worth doing roughly at twenty-four.

The third change is the one people feel without naming. At an illustrative 11% a year — an assumption, not a forecast — a rupee saved thirty years before retirement becomes ₹22.89, because 1.11 raised to the thirtieth power is 22.89. The same rupee saved twelve years out becomes ₹3.50. Late money does roughly a seventh as much work as early money, so a high income arriving in this decade is not simply a windfall. Part of what it is doing is paying for the shorter runway underneath it.

Which cuts both ways, and the second half is usually left out. Income for many people is at or near its highest in this decade — a pattern plenty of careers follow and plenty do not — and some of the claims that made saving hard earlier may have ended: a course paid for, a loan closed, a deposit found. The amount you can add per year is larger than it has ever been at exactly the point each rupee does less. Neither fact cancels the other. Together they explain why this decade decides most retirement outcomes and why almost nothing about it is obvious from the outside.

What risk means when the recovery time is short

The usual conclusion drawn from a shorter runway is that a portfolio should hold less in equity as its owner ages. That is a category error, and an expensive one. Money held by a forty-seven-year-old is not spent at sixty; the last rupee of it is spent in their eighties. Most of that portfolio still has a holding period of thirty or forty years. Age is not a horizon — the horizon belongs to the money, not to the person holding it.

What genuinely changed is that several large sums have acquired dates. A fee, a deposit, a planned gap between roles, a relocation, a parent's care. Risk in this decade is not the chance that a holding falls; it is the chance of being made to sell it on a date you did not choose. A fall does not have to be permanent to do damage. It only has to coincide with the date attached to it.

Read the table as one instruction rather than four: match the holding to the date, not to the birthday. That is the subject of asset allocation and goal-based investing, and it produces a different portfolio from the age-based rule — often a more conservative one for the next three years and a more aggressive one for the money that will not be touched for twenty.

One consequence is specific to the size of a forties portfolio rather than to the age of its owner. At thirty a rebalance was a click. At a corpus of a crore it is a realisation event: under the Income-tax Act 2025, long-term gains on equity and equity-oriented fund units are taxed under s.198 at 12.5% above an annual ₹1.25 lakh, and that applies under either regime. Directing new contributions at whatever is underweight shifts the ratio without a sale, which is why the same rebalance costs nothing at one portfolio size and something at another.

The decade your largest asset changes identity

At twenty-eight the biggest thing you own is not in any account. It is the earnings you have not received yet, and no statement reports it. Somewhere in the forties that stops being true, and the shift explains several things that otherwise look like contradictions.

Put a number on it, with inputs to substitute. Take earnings of ₹15 lakh a year and discount them at an illustrative 5% after inflation. Thirty years of that stream is worth ₹2.31 crore today, because thirty years discounted at 5% comes to 15.37 times one year of it — the multiplier has a name, the annuity factor. Twelve years of the same stream is worth ₹1.33 crore, the factor having fallen to 8.86. The two curves cross at the point where the portfolio passes the discounted value of what is left to earn, and the second of those two numbers falls every year whether or not the first one rises — which makes the crossing a matter of arithmetic rather than of occupation.

Three things follow, and each is mechanical rather than a matter of taste. The first is about protection. Cover on a life exists to replace what has not yet been earned, so the gap it fills shrinks as the corpus grows and the earning years run down — which is why a sum assured chosen at thirty-two is usually the wrong number by forty-seven, in one direction or the other. If anyone depends on your income, or a loan outlives you, the arithmetic is in the term insurance guide; the point here is only that the input has moved.

The second is about which shock now dominates. In your thirties the largest threat to net worth was losing the income, because the income was the larger asset. Increasingly it is a fall in the portfolio, because the portfolio is. That is the honest content of "less time to recover" — not that markets became more dangerous, but that a given percentage fall now applies to a bigger share of everything you have.

The third is the counterweight, and it is the subject of the deferral section below. Human capital has shrunk but it has not gone, and it remains the single most powerful variable available — more powerful, at this point, than any decision about what to hold.

When more than one generation has a claim

Start by refusing the assumption. Some people in their forties support a parent; some support a child; some support both, which is the case the phrase "sandwich decade" was coined for; some support a sibling, a partner's family or a dependent adult; and some support nobody and are quietly told they have it easy. The constraint below applies where it applies, and the arithmetic is the same whoever is at the other end of it.

The useful split is not between generations. It is between claims that have a date and claims that do not. A fee due in year four is datable: it has a size, a deadline, and it can be funded on a schedule — the mechanics are in sinking funds. A health event has neither a date nor a ceiling. It cannot be scheduled, and no allocation makes it safe.

The common inversion is to treat these backwards: to buy a savings-linked product for the datable claim, whose date and size are already known, and to self-fund the undatable one by holding cash "in case". Insurance exists precisely for the claims you cannot put a date on. Undatable claims want cover and liquidity, not schedules.

Where an older parent is the person at the other end, three mechanics decide most of the outcome and all three are about timing rather than price. Cover bought after a diagnosis does not pay for that diagnosis: a pre-existing condition can carry a waiting period of up to 36 months under the IRDAI product regulations. Cover that has run for 60 months reaches what the regulations call a moratorium, and that clock starts again on any increase in the sum insured. And for indemnity individual cover on someone aged sixty or above, premium revision is capped at 10% per annum — which matters because open-ended renewal pricing is the fear that most often stops the policy being bought at all. Whether the cover belongs in a shared policy or a separate one is a structural question, and what a claim actually involves is in medical emergency planning.

Two further points that get missed. Employer group cover is not yours; whether it survives a job change or a retirement, and on what terms it converts, is a question worth asking of the policy document rather than assuming either way. And some claims have no insurance product at all — an adult who needs ongoing support is a permanent reduction in the surplus, and the specific error is budgeting it as temporary for the eighth year running.

One last mechanic that only bites at this decade's balances. A sinking fund for a large datable claim can exceed ₹5 lakh, which is the deposit insurance cover per depositor per bank, principal and interest together. At twenty-five that ceiling sat above anything you were likely to hold. Here it does not, and how the cover aggregates becomes a live question rather than a curiosity. The emergency fund sits underneath all of it, and what it has to survive in this decade is set out in the job-loss plan — a more specialised role has fewer positions open at any moment, which is structural rather than statistical.

The one loan you cannot take

Here is the specific error, named precisely enough to recognise. A large education bill arrives — a degree, a professional course, a move abroad, for a child or for yourself — and it is met by withdrawing from retirement savings, or by pausing contributions to them, on the reasoning that the education is urgent and retirement is distant. Both halves of that reasoning are true. The conclusion does not follow.

The asymmetry is mechanical and it is the whole argument. A lender will write a loan against a degree, at a stated rate, over a stated number of years, with a repayment schedule you can see before you sign it. Against a retirement corpus you failed to build, no lender writes that loan. There is no instrument, at any price. That is not a claim about which is more important; it is a claim about which of the two shortfalls has a market in it.

Put the withdrawal in the units it is actually paid in. Take ₹20 lakh withdrawn at forty-eight, an illustrative real return of 3%, and retirement at sixty funding thirty years. By sixty that ₹20 lakh would have been ₹28.5 lakh in today's money, since 1.03 to the twelfth is 1.4258. Divide by the thirty-year annuity factor at a real 3%, which is 19.60, and the withdrawal costs ₹1.45 lakh a year of retirement income — every year, for the whole of it. The number that matters is not the ₹20 lakh. It is the income the ₹20 lakh was going to pay.

Against that, the loan. ₹20 lakh over ten years at an illustrative 10% is about ₹26,430 a month, or ₹3.17 lakh a year, and in the eleventh year it stops. Those two figures are in different units — the repayments are in the rupees of the years they are paid, the ₹1.45 lakh is in today's money — so they must not be netted against each other. Read their shapes instead. One claim is bounded, dated, transferable and payable while you are still earning. The other is permanent, arrives when you are not earning, and cannot be re-borrowed.

Now the version nobody counts, because it feels like restraint rather than spending. Pausing a ₹50,000 monthly contribution for four years from forty-eight puts ₹24 lakh less into the corpus. At the same illustrative real 3%, those four instalments would have grown to ₹31.8 lakh by sixty — six lakh multiplied by 1.03 raised to the eleventh, tenth, ninth and eighth powers — which is ₹1.62 lakh a year of retirement income. The pause costs more than the withdrawal did, and it is easier to make precisely because nothing leaves the account.

The trade-off, since nothing here is free. A loan is a fixed claim on an income that may not be secure, it may need a co-borrower, and somebody carries it for a decade; the structure is in the education loan article. Where a household has genuine surplus, paying from cash flow avoids the interest entirely and is not the error described here. Interest relief on such a loan sits in s.129 of the Income-tax Act 2025 and belongs to the old regime, so a reader on the default new regime under s.202 gets none of it — which is the kind of thing that changes the comparison and is worth checking against the regime comparison rather than assumed. What the arithmetic rules out is narrower than "do not borrow" and narrower than "do not pay cash". It rules out treating the corpus as the cheapest of the three, when it is the only one with no replacement.

Why the last working years do the most work

The most powerful variable in this decade is not the return assumption and not the fund selection. It is the year you stop, and the reason is that moving it acts on three quantities at once instead of one.

Take the illustrative case, kept entirely in today's money so nothing gets mixed: a corpus of ₹1.5 crore at fifty-nine, contributions of ₹9 lakh a year, spending of ₹9 lakh a year in retirement, a real return of 3% and a retirement that has to be funded to about ninety. That last assumption is the load-bearing one and it is stated deliberately — the horizon is anchored to longevity, not to the leaving date, so stopping a year later is one fewer year to fund, not the same thirty years pushed outward.

One additional year closed 65% of the gap. Two years closed it completely and left a margin. Nothing in that table required a better return, a better fund or a better decision about anything — the year added a real 3% to the largest balance of your life, added a full year of contributions on top, and removed a year of withdrawals from the far end. Three effects, one decision. The annuity factors are 19.60, 19.19 and 18.76 for thirty, twenty-nine and twenty-eight years at a real 3%, and the corpus line is the previous row multiplied by 1.03 plus ₹9 lakh.

Two honesty notes on those figures. The real rate here is a round 3% for legibility; the properly derived version divides rather than subtracts, and the corpus article shows the shortcut is worth about 2% on a thirty-year number, against the 49% that mixing units costs. And the power of the lever depends on the corpus already being large, which is exactly why it is a forties-and-fifties instrument and not one available at thirty.

Run the same arithmetic backwards and it explains what a forced exit actually costs. An involuntary stop at fifty-seven does not remove one year of earnings. It removes the highest-yielding years in the whole sequence, and it removes them from the end, where the balance they were acting on is biggest. That is why cover against illness and disability is a portfolio question in this decade rather than a domestic one — the mechanics are in critical illness and disability cover.

Which forces one more piece of honesty. Working longer is not universally available. Health, sector, employer decisions and caring responsibilities all remove it, and they tend to remove it at short notice. A plan that depends on three extra years is a plan with an unhedged assumption in it, so the deferral lever is best counted as an option, not a plan — valuable if it can be exercised, and disastrous if the whole shortfall was resting on it.

Four errors that only become available in this decade

Each of these needs a large portfolio, a short runway or both. None of them is possible at twenty-five, which is why generic advice does not name them.

1. Catch-up gambling. A gap gets discovered at forty-six and is closed by taking concentrated risk — a single stock, a leveraged position, a high-return promise. The mechanism is worth stating plainly: raising the return assumption changes the spreadsheet, raising actual risk changes the distribution of outcomes, and with no recovery time left the bad half of that distribution has nowhere to go. The levers that genuinely move the answer are contribution, horizon and target spending, and the contribution arithmetic is where to start.
2. The unreviewed accumulation. Twenty years of holdings collect a long tail: schemes bought for a reason that expired, policies bought for a tax break under a regime you no longer use, an employer's stock that has quietly become a concentration. The size is what makes this expensive now — the same neglect cost almost nothing on a two-lakh portfolio. What a tax break is worth depends on which regime you are in, which is a live question every year.
3. Documents describing an earlier life. Nominations naming a parent who has died, a will written before a household changed, a beneficiary who is no longer the intended one, and assets nobody else knows exist. The fix is clerical and the consequence of skipping it is not — see nomination, a will and estate planning.
4. Reading the surplus as permanent. A high income with a temporary absence of claims invites a rise in fixed costs, and fixed costs are what make a forced exit expensive later. What is specific to this decade is that a fixed cost added now is also capitalised: on the illustrative real 3% and thirty-year horizon used above, ₹1 lakh a year of permanently higher spending raises the corpus you have to reach by 19.6 times that, or ₹19.6 lakh — and it has to be found from a runway that no longer has thirty years in it. That ratchet is the subject of lifestyle inflation, and it does its worst damage in exactly the decade the surplus looks safest.

None of the four is a failure of discipline. Each is what happens when a portfolio and a set of obligations both grow faster than the habits that were built around smaller versions of them.

Testing the plan against the record rather than the average

Everything above runs on a constant real return, which is a modelling convenience and not a property of any portfolio. That gap matters more here than it did earlier: a plan that reaches its corpus on the average path may not reach it from the start dates that actually occurred, and with a short runway there is no second attempt.

FNOTrader's Mutual Funds app runs on the full history of published daily per-unit prices — net asset values, or NAVs — from the Association of Mutual Funds in India, around 34 million rows of them. It simulates a contribution schedule on any scheme and period and reports the return on cashflows that land on irregular dates — XIRR — along with invested against value, the deepest fall from a previous peak along the way, and rolling-return distributions across every available start date. Two tests follow directly from the arithmetic above: whether the remaining instalments reach the corpus using the worst historical window rather than the mean, and what a drawdown of that depth would have done to a sum that had a date three years away.

Historical outcomes describe what happened, not what will happen; past performance does not indicate future results. FNOTrader is not a SEBI-registered investment adviser and nothing here is advice. What suits a particular household depends on facts an article does not have, and the drawdown behaviour of a withdrawal phase is a separate subject covered in withdrawal rates and sequence risk.

Common questions

Continue reading

More in Money by Life Stage · App: Mutual Funds · Definitions: glossary · Free tools: calculators · All: every article