The Retirement Withdrawal Problem: How Much Can You Actually Spend?

Saving is only half the story

Most retirement planning energy goes into one question: how big a corpus do I need? You pick a number, work backwards into a monthly SIP, and feel reassured once the corpus projection crosses that number.

But reaching the number is not the finish line. The moment you retire, a new and much trickier problem begins: how much can you withdraw every year without running out of money before you run out of life? Get this wrong and even a large, well-earned corpus can empty out faster than you expect.

This is the withdrawal problem, and it deserves as much attention as the saving problem — arguably more, because mistakes here are much harder to undo.

Why the order of returns matters, not just the average

Here is something that surprises a lot of people: two retirees can earn the exact same average return over 25 years and end up with wildly different outcomes, just because of the order in which the good and bad years happened. This is called sequence-of-returns risk.

The reason is simple once you see it. While you are still working and adding money every month, a market crash is actually fine, even helpful — you buy more units at lower prices, and the market eventually recovers. But once you retire and start withdrawing, you are selling units to fund your expenses. If a crash hits early in retirement, you are forced to sell more units at depressed prices just to get the same amount of cash. Those units are gone for good — they can't participate in the recovery that follows, because you've already sold them.

A downturn in year two of retirement can permanently dent a corpus in a way that the same downturn in year twenty never could, even though the long-term average return is identical in both cases. This is why retirement withdrawal planning can't just look at an average return assumption — it has to account for the risk of bad luck showing up early.

The idea of a sustainable withdrawal rate

To manage this risk, planners use the idea of a sustainable withdrawal rate — the percentage of your corpus you withdraw in the first year of retirement, which you then increase with inflation each year after. Pick a rate that is too high, and there is a meaningful chance the corpus runs dry while you are still around to need it. Pick a rate that is comfortably low, and the corpus is far more likely to outlast you.

You may have heard of the "4% rule" — a rough starting reference that came out of research on US markets and inflation. It is a useful mental anchor, but it does not transfer cleanly to India. Indian inflation has historically run higher and been less predictable than US inflation, and Indian equity and debt markets have their own return and volatility patterns. That combination generally argues for starting more conservatively than 4% in an Indian context, especially for a long retirement of 25-30+ years, rather than assuming the same number applies here.

Think of a withdrawal rate less as a fixed rule and more as a planning input you stress-test — you want to see how your specific corpus, spending, and time horizon behave under different rates and different market conditions, rather than picking one number and hoping for the best.

Other income reduces what the corpus has to carry

Not every rupee of retirement spending has to come out of your investment corpus. If you have a pension, rental income, or any other steady inflow, that income offsets your expenses first — and the corpus only needs to fund the gap that is left over.

This matters more than it sounds, because withdrawal sustainability is highly sensitive to how much you are pulling out each year. A pension that covers even a third or half of your monthly needs can substantially lower the effective withdrawal rate on your corpus, which in turn meaningfully improves how long it is likely to last. When you are estimating how much corpus you actually need, it is worth being precise about any offsetting income rather than ignoring it and over-building (or worse, under-building) your target.

A simple illustration

Say someone retires with a corpus of ₹2 crore and estimated monthly expenses of ₹1 lakh, or ₹12 lakh a year. If they have no other income, the corpus has to fund the full ₹12 lakh — a withdrawal rate of 6% in the first year, which is on the higher, riskier end for a long Indian retirement.

Now suppose the same person also receives ₹40,000 a month from a pension or rental property, or ₹4.8 lakh a year. The corpus now only needs to cover the remaining ₹7.2 lakh a year — a withdrawal rate of 3.6%, a meaningfully safer starting point. Same corpus, same lifestyle, very different odds of the money lasting — purely because of how much of the burden the corpus alone is carrying.

This is a simplified illustration and ignores things like tax, changing expenses, and real market sequences, but it shows the core lever clearly: the withdrawal rate, not just the corpus size, is what determines how long your money lasts.

Try it with your own numbers

These ideas are much easier to internalise when you can see them play out with your own corpus, expenses, and any pension income, rather than just reading about them in the abstract.

finkit's Retirement SWP calculator lets you see how long a corpus is likely to last for a given monthly withdrawal, with an optional pension offset, so you can test different withdrawal rates against your own numbers. The SWP Calculator is useful more generally for modelling any systematic withdrawal plan, and the Can I Retire? calculator helps you work backwards from your expenses to check whether your current corpus and savings plan are actually on track. All three are free, run entirely in your browser, and do not ask you to share any personal financial data to use them.