Healthcare Costs in Retirement

Most retirement budgets are built from things people can picture: groceries, electricity, travel, the occasional gift. Healthcare tends to get a single line and a round number, usually chosen because it looks reasonable rather than because anything was calculated. It is the one line in the budget that reliably grows in two directions at once — the price of each unit of care goes up faster than everything else, and the number of units you consume goes up as you age.

That double compounding is what makes it worth separating out. Averaged into the rest of the budget, it disappears. Modelled on its own, it usually turns out to be one of the largest single claims on a retirement corpus.

The two rates that matter, and why they are not the same

India's headline CPI has been running inside the RBI's 2-6% tolerance band, with the 4% target retained for the five years to March 2031. Private medical costs have behaved nothing like that. Estimates for 2025-26 have clustered in the 12-14% range, and NITI Aayog has cited an average near 14% over the preceding five years. These are industry and consultancy estimates rather than an official index, so the precise figure is arguable — but the gap between medical and general inflation is not.

Consider what that gap does over time. A treatment costing ₹3 lakh today, inflated at 5%, costs roughly ₹8 lakh in twenty years. At 12%, the same treatment costs around ₹29 lakh. Same procedure, same twenty years, a difference of more than ₹20 lakh produced entirely by which rate you assumed. Nothing else in a retirement model is this sensitive to a single input, which is precisely why folding healthcare into a blended household rate quietly understates it.

Our inflation calculator lets you set a different rate on the medical line than on the rest of the budget, which is the minimum version of this exercise. The healthcare cost calculator takes it further and treats the medical reserve as its own corpus.

Volume rises with age, independently of price

Price is only half of it. A 62-year-old and an 82-year-old are not buying the same quantity of healthcare at different prices — they are buying different quantities. More consultations, more routine diagnostics, more maintenance medication taken indefinitely rather than for a course, more dental and vision work, and a rising probability of at least one significant intervention.

Then there is the category people leave out entirely: assisted living, home nursing, attendant care. It is the least predictable line — many people never need it, some need it for a decade — and it is largely outside the scope of a standard hospitalisation policy. A plan that assumes total spending drifts downwards through retirement is usually right about travel and dining and wrong about this, and the two do not offset each other in size.

What insurance covers, and what it leaves you holding

A health policy is good at one specific thing: absorbing a large, sudden, hospitalisation-shaped cost that would otherwise force a distress withdrawal from the corpus at the worst possible moment. That is a real and significant benefit, and it is the reason cover matters more in retirement than during working years, when a salary can absorb shocks.

What a policy generally does not absorb is the steady drip. Outpatient consultations, diagnostics, dental, vision, physiotherapy, ongoing medication, and whatever sits inside the policy as room-rent caps, disease sub-limits, co-payment clauses and deductibles — those come out of the corpus regardless. Nor does the policy cover its own premium, which is a recurring expense that grows every single year of retirement.

There is also the adequacy question. A ₹10 lakh sum insured looks generous today. At 12% medical inflation, the procedures it comfortably covers today cost close to ₹31 lakh in a decade and roughly ₹96 lakh in two. Cover chosen once at retirement and never revisited becomes progressively less useful in exactly the years it is most likely to be claimed against — which is an argument for treating sum insured as something to review periodically rather than a decision made once.

Premiums are their own inflating line

Two things push senior premiums up: the underlying medical inflation, and age-band repricing as you cross each threshold. Annual premiums in the tens of thousands of rupees per person are ordinary at older ages, and for a couple that is a meaningful recurring cost sitting inside a budget that has no salary behind it.

There is some regulatory constraint. IRDAI has directed insurers not to raise renewal premiums for policyholders aged 60 and above by more than 10% a year without prior approval, which caps the year-on-year shock for someone already holding a policy. It does not do much for someone shopping for fresh cover at 65, where the entry price already reflects the age band — one of the reasons continuity of cover is worth more than it appears on a premium comparison.

On tax, the position has moved recently. GST on individual health insurance policies, including family floaters, was reduced to nil with effect from 22 September 2025, so a retail premium quoted now does not carry the earlier 18% addition; group policies were not covered by that change. Section 80D continues to allow a deduction on premiums, at a higher ceiling where the insured is a senior citizen, but only under the old tax regime — it is not available under the default new regime. Both points are current as of this writing and have changed more than once; confirm them for the relevant year before relying on either.

The government backstop, and its limits

Since October 2024, the Ayushman Vay Vandana card has extended AB PM-JAY cover of ₹5 lakh a year to everyone aged 70 and above regardless of income, with tens of lakhs of cards issued. For a hospitalisation inside the empanelled network it is a genuine backstop, and it did not exist for the current generation of retirees when they started planning.

Its limits are equally worth stating. It begins at 70, which leaves the sixties to be funded some other way. It operates within a defined network and package-rate structure, so it does not map onto every treatment at every hospital. And ₹5 lakh is a nominal figure — at any plausible medical inflation rate, its real value in the 2040s is a fraction of what it buys today. It is sensibly treated as something that reduces the tail risk, not as a line item that funds the medical budget.

Building it into the plan

The mechanics are less complicated than the topic sounds. Pull the medical line out of the household budget and inflate it at its own rate rather than the blended one. Add the insurance premium as a separate, separately inflating expense — it is a certainty, not a contingency. Assume the quantity of care rises through retirement rather than staying flat, even if only crudely, by stepping the medical budget up at defined ages. Then run the whole thing through the retirement corpus calculator and compare the required number against the version where healthcare was averaged in with everything else. The difference between those two figures is what the shortcut was costing.

Two structural points fall out of the arithmetic. Liquidity matters more for this line than for any other, because medical costs arrive on their own schedule and often during exactly the market conditions when you least want to sell — an accessible reserve is what stops a health event from becoming a portfolio event. And the medical portion of the corpus faces the longest, fastest inflation of anything in the plan, which argues against parking all of it in instruments that barely keep pace with general prices, at the cost of accepting some volatility on money you might need at short notice. That is a trade-off with no clean answer, and where it lands depends on how much of the rest of the corpus is already liquid.

Assume you will be wrong, and size the margin

Nobody knows their own medical trajectory, and no assumption you choose will turn out to be the right one. That is not a reason to skip the exercise — it is the reason to do it at several rates rather than one. Run the plan at 8% medical inflation, then at 12%, then at 14%, and look at the spread. If the plan holds across the range, the margin is adequate. If it only holds at the bottom, the plan is not conservative; it is just quietly assuming the most favourable input.

Healthcare is not the line that makes retirement planning hopeless. It is the line that makes vague planning expensive, and it is the one most improved by being separated out and looked at directly.