Your Plan Ends Before You Do

Every rule of thumb in Indian personal finance descends from one unstated ratio.

Start earning at 25. Stop at 60. Draw for 20 years. Roughly two working years fund every retired year, which is demanding but manageable, and it is the shape your client's parents planned around.

One thing has already happened to that shape. The horizon has stretched to 100, and on that horizon the ratio has quietly crossed one.

A second thing might happen to it. For some clients, the earning window may not stay open until 60.

The first is arithmetic. The second is a scenario. This edition treats them differently, and so should a plan.

How long the money has to last

Set the two halves of a working life side by side and count the years in each.

Under the old shape, a client earned from 25 to 60 and drew for twenty years. Thirty-five years against twenty. The earning half was nearly twice the length of the retired one, and that is the arithmetic every conventional Indian planning rule was built on.

Most plans are a little more careful and run to 85. That makes it 35 against 25. Still comfortable.

Now hold the career exactly where it is and extend the horizon to 100. The same 35 earning years, and 40 years after earning stops.

The retired half is now the longer one. The client still works to 60. Nothing has gone wrong.

Read down the two columns and find the row where they cross.

In the first row, the earning years lead by fifteen. By the second, the lead is down to ten. By the third, the client is behind by five, and every row after that widens the gap.

Row three is where most clients already are. It requires no view on artificial intelligence, no assumption about early retirement, and no forecast of any kind. It happens on its own, from the horizon alone, and it has already happened.

The rows below it are different. They assume the earning window closes before 60, which is a scenario rather than a fact, and it needs handling with more care than most commentary gives it.

That is what the next section is about.

Recommended for you

Readers also explored

Blogs

World GDP Breakdown 2025: Who Powers the Global Economy?

8 min read

Blogs

Nifty 50 Companies List 2025 : Top 50 Stocks in India

7 min read

Where AI comes into this

The rows below the third one assume the earning window closes early. That needs justifying, and the justification is narrower than most commentary offers.

Careers are not currently ending at 45. There is no evidence they are, and anyone saying otherwise is forecasting rather than reporting. Displacement of experienced mid-career workers has not been demonstrated anywhere.

But the mechanism that would shorten a window does not require anyone to lose a job.

An earning window closes when accumulated experience stops commanding a premium. That is the proposition a client's whole plan rests on: that the 45-year-old is worth more than the 30-year-old because of what sits between them, and that the gap widens through the fifties.

AI puts pressure on exactly that proposition. When a capability that took fifteen years to build becomes available to someone with two years and the right tools, the premium attached to those fifteen years compresses. The worker keeps the job. The salary stops climbing.

That is already visible in India, and our work on the Indian AI workforce shows the shape of it. AI and data roles pay 20 to 40% more than comparable traditional IT roles, with experienced professionals at ₹35 to 70 lakh against ₹23 to 58 lakh on plateaued tracks. India leads the world in generative AI course enrolments while ranking 89th of 109 countries on measured proficiency.

So the premium is moving faster than the workforce is acquiring it. A client on the wrong side of that does not become unemployed. They plateau, and a plateau at 45 has the same effect on a plan as an exit at 55, because the years that were supposed to carry the heaviest contributions carry the same ones as a decade earlier.

It is also worth knowing that a shortened window is not hypothetical in India. At Tata Consultancy Services, 1.1% of the Indian workforce is over 50, against 21.8% at the same company in the United States and Europe. Infosys sits at 2.5%. Tech Mahindra cut its retirement age from 58 to 55 for junior bands in 2018, and all of it predates generative AI entirely.

A twenty-year professional career already exists in a large, formal, high-paying Indian industry, for reasons that have nothing to do with technology. If it can happen there without AI, it is not exotic to model it happening elsewhere with AI.

How to use this with a client. Not as a prediction. A plan built to hold if the earning window closes at 50 also holds if it closes at 60. The reverse is not true. That asymmetry is the entire argument, and the client needs to accept nothing about artificial intelligence to accept it.

What a shorter window costs

Ask a client what an early exit costs, and they will say the salary they did not earn.

The corpus rises 21% across those rows. The monthly saving needed rises 150% because a shorter window lengthens what must be funded and shortens the time available to fund it.

In the room. Run the plan at 60, 55 and 50, and show the required monthly savings rather than the terminal corpus. Frame the lower rows as insurance, not prediction. The client is not being asked to believe their career ends at 55. They are being asked whether they would rather find out at 42 or at 55.

The squeeze

The above table assumes those savings are available. For a household in its forties, they are not, because the same income funds obligations in three directions.

• Downward, to the child. 

Every education figure quoted today is a sticker price.

A school education and an IIT degree together come to about ₹30 lakh. The same child through an elite school and a US MBA is nearly ₹2.5 crore.

Both are choices ordinary Indian families make. The gap between them is larger than the retirement corpus most of these households are working toward.

And these are today's prices. Fees compound at 11 to 13% a year, so a family paying six or ten years from now pays roughly double.

Which makes education less a line item with a plausible average and more a single decision that moves the retirement position by more than the retirement plan does. It usually gets taken on grounds that have nothing to do with money, and it rarely gets discussed as a retirement decision.

• Upward, to parents

The second claim is the one households budget as a flat monthly transfer, and it is the least flat thing on the list.

Source: Aon, 1 Finance Research. Note: Starting premium is a broker quote for a floater covering a couple aged 72 and 71. Escalation at the 11.5% India medical trend rate for 2026 from Aon. Medical trend rates describe employee medical plan costs and are not a CPI measure of household medical inflation.

Nothing about the parents' health has to deteriorate for the premium to reach ₹6.20 lakh. The trend rate alone does it.

Premiums are also only the visible part. Out-of-pocket costs, housing support and the shocks an older parent cannot absorb sit on top, and none of them can be put on a schedule.

• Sideways, to the adult child. 

The third claim is the one nobody budgets at all, because it is supposed to have ended.

In the latest labour force survey, unemployment was 3.1% for those aged 15 and above, 9.9% for youth aged 15 to 29, and 6.5% among those educated to secondary level or higher. The gradient runs the wrong way for this readership: unemployment is markedly higher among the more educated.

Which means the young adult whose education consumed a share of the retirement corpus is also the one likely to spend a stretch without income afterwards. The household funds the degree, then funds the period after it.

India does not publish how long that period lasts. It is the single most useful missing statistic in this section.

Why the timing matters more than the total

None of the three claims is unmanageable alone. The problem is that they do not arrive evenly, and plans are tested against annual averages.

A worked example makes the shape visible. Take a household with post-tax income of ₹50 lakh, one child, and two parents in their early seventies with no pension. They save ₹6.57 lakh a year for retirement, which is 13% of income.

Three moments in that household's life, each showing what it commits in that year.

The master's figure is the one that needs explaining. A two-year US master's costs about ₹105 lakh at today's prices, or ₹52.5 lakh a year. Fee inflation and the rupee together take that to ₹124.3 lakh a year by the time it is actually paid.

For a decade this household looks like a planning success. Twenty per cent of income going out, never a missed contribution, school fees absorbed without strain.

Then one year arrives in which committed claims exceed the entire year's post-tax income by half again.

Four of those claims have deadlines. The school, the university, the insurer, and the education loan if there is one. Exactly one does not.

So the retirement contribution stops, usually for the two years the child is away. That is not a hypothetical. It is what happens when the arithmetic of that column resolves itself.

The amount missed is identical in every row. What changes is how many years are left to replace it.

A pause at 45 is a nuisance. The same pause at 52 needs a permanent 27% increase in saving for the rest of the working life, in the decade when income is likely to flatten and when, on the earlier argument, the earning window may be narrowing anyway.

In the room. Put education, parent support and retirement contributions on one timeline through the client's forties and fifties. You are looking for the single worst year, not the average. That year is knowable a decade in advance and rarely modelled.

What each decade costs

Assume the client gets through all of it and retires at 60. Most plans model what follows as a flat real withdrawal for forty years.

Two forces say otherwise, and they run in opposite directions. Lifestyle spending falls at roughly 1% a year in real terms as travel and discretionary consumption soften. Health spending rises at about 6.5% a year in real terms, being India's 11.5% medical trend against general inflation near 5%.

Lifestyle spending does fall, exactly as expected. Across forty years it drops ₹1.49 lakh. Health rises ₹11.53 lakh over the same period and buries it.

So real spending does not decline through retirement. It more than doubles, and health goes from a quarter of the bill to three-quarters. The client is funding one kind of life at 62 and something else entirely at 92.

Which is why the earlier number was wrong

The corpus in the first table assumed a flat ₹12 lakh a year for forty years.

Reprice the same retirement on the profile above, from the same starting point, and the requirement rises by 40%. Required monthly savings at a 60 exit goes from about ₹55,000 to ₹76,000.

For a household already saving ₹55,000 and feeling comfortable, that is the entire problem, and it sits in an input nobody revisits.

So what do you tell them?

A 40% shortfall is not a savings target. It is the wrong conversation.

Working longer is the strongest single lever, and it does not get there alone. Nor does any other. Together they take the requirement below what most of these households already save.

That is the conversation. Not "save more," which invites a client to nod and change nothing, but three specific changes, none sufficient alone, of which they need roughly all.

The third is worth saying out loud. Property is usually the only asset large enough to matter at this scale. If the plan only works with the home in play, the plan already assumes it, and that is better established at 42 than at 68.

What to do with this

• Reprice the spending assumption first. A flat real withdrawal understates the corpus by 40%, and every other figure in the plan derives from it. Nothing else on this list is worth as much.
• Then stress the earning window. Three plans at 60, 55 and 50, with required monthly savings as the output.
• Show the client the education ladder before they choose. The gap between the cheapest path and the dearest is larger than their retirement target, and that conversation belongs at 42, not 52.
• Find the collision year. One timeline, three claim streams, look for the peak.
• Present the levers as a set, never one at a time. No single change is sufficient, and offering them individually invites the client to reject each on its own terms.