What's changed
When we change an assumption, your number changes with it. That shouldn't happen quietly, so every update that moves the maths is recorded here — including the ones that make our own earlier claims look worse.
September 2026
equity assumption lowered · corpus calculator addedWhy your number may have changed
We lowered the assumed long-term return on equity from 12% to 10% a year. It applies to the equity and direct-stocks defaults on the Retirement page, the equity assumption inside the 3-bucket withdrawal strategy, and the average return used by the Stress Test's Monte Carlo simulation.
Two reasons. 12% was an optimistic reading of long-run Indian equity returns in nominal terms. More importantly, it made our own strategy comparison unfair: the 3-bucket strategy was being credited with 12% equity while the Simple strategy ran at a flat 8%, so part of what looked like a better strategy was really just more risk.
What it does to your numbers: the sustainable monthly withdrawal under the 3-bucket strategy falls by roughly 10%. If you enter your savings split by instrument, your projected corpus is lower and the monthly SIP we suggest is correspondingly higher. Plans that use the single “total saved” field with its own growth rate are unaffected.
If you already have a saved plan
Your plan keeps the assumptions you saved with it. If you turned on Keep on this device, or you load a .kal.json file you exported before this update, you will still see 12% — we don't rewrite your saved numbers behind your back. That does mean two people entering identical inputs can now get different answers, depending on when their plan was made.
To move to 10%, either change it yourself under Assumptions → Equity return, or use Reset to start from the current defaults. One exception worth knowing: the Stress Test's Monte Carlo average return isn't stored with your plan, so it uses 10% even if the rest of your plan is on 12%.
A claim we corrected
We used to say the 3-bucket strategy produced a dramatically larger ending balance than a simple portfolio, on the strength of a “1–4 percentage point return edge.” That was overstated, and the reason is the assumption above — we were comparing a bucket portfolio holding more equity against a more conservative flat rate, then attributing the gap to the strategy.
Corrected: at matched risk (both approaches now work out to roughly 43% equity), buckets are modestly ahead on smooth average returns — about 6% more sustainable monthly spending, not a transformation. The genuine case for buckets was never the return. It's that a crash in your first few retirement years can't force you to sell equity to fund that year's spending. The full explainer now says so plainly.
Corrections to the maths
- The required-corpus search now verifies its own answer. It searched between zero and an upper limit that was assumed, never checked, to be large enough. With punishing assumptions — high inflation, a low post-retirement return, a very long horizon — that limit could be too small, and the figure returned was one that did not actually fund the plan. The search now raises its limit until it finds one that survives. If you use assumptions near those extremes, your required corpus may now be higher, and right.
- The monthly SIP search no longer under-shoots on large Big Spend goals. A big spend late in your saving years can drain the corpus faster than the old search allowed for, so the recommended monthly amount could come back too low to actually reach your target.
New
- Work out the corpus you need — a third mode that runs the calculation backwards: tell it the monthly income you want and when it should start, and it solves for the smallest corpus that funds it.
- A full 3-bucket explainer — what the buckets are, the risk they remove, the rules for moving money between them, and what the approach costs you.
- Simple vs 3-bucket, side by side — on your own numbers, in average markets and through an early crash. It shows in both strategies, because the people most likely to benefit from buckets are the ones who never tried the toggle.
- A chart of how the buckets drain and refill across the whole plan, rather than just their opening sizes.
Removed and fixed
- “Monthly take-home income” is gone from the Retirement page. It fed no calculation. We checked by running the plan across incomes from ₹1 to ₹1 crore: the required corpus did not move by a single rupee. A field that looks like it matters and doesn't is worse than no field.
- The two age sliders now share one scale. “Your age” and the age withdrawals start were drawn on different ranges, so — for example — 43 and 60 could land on the same spot on the track.
- PDF and Excel exports understand the new corpus mode instead of describing it as a plan that runs out immediately.