Should you stop your SIP when the market falls?

Sometimes, but much less often than people expect. We ran every combination of sell trigger (5%–40% off the peak), buy-back level (10%–60%) and share moved (20%–100%): 12,060 rules. Against a plain SIP returning 20.97% XIRR, 2,809 of them (23.3%) came out ahead. The median rule lost 0.78% a year.

So three out of four timing rules did worse than doing nothing, and there was no way to know in advance which group your rule was in.

Run the sweep yourself →

Most of the winners stop winning. Switch on Indian capital-gains tax (LTCG at 12.5% with the yearly exemption, STCG at 20%, liquid-fund interest at your slab rate) and the same 12,060 rules give a very different result:

Most articles on SIP timing leave this number out, because it requires working out the tax on every switch.

The history says keep going, even from the worst possible starting point. A SIP begun in January 2008, one month before the index fell 69% from its peak, still compounded at 22.07% a year. The falling months were when the SIP bought cheaply. Stopping would have skipped those purchases.

Open the 2008-peak backtest →

Waiting usually costs more than the crash. The months you sit out tend to be the cheap ones. You can test this in the lab: set a sell trigger and see how many months the rule spent in a liquid fund. The 20/35/70 rule sat out 61 of 257 months and still ended behind a plain SIP.

This one holds up, slightly. A buy-the-dip SIP (1.5× the monthly amount past a 10% fall, 2× past 20%, 3× past 30%) returned 20.52% after tax against 20.06% for the flat SIP. The gain is real but small, about half a percent, and it needs spare cash ready every time.

Run buy the dip →

Barely, per rupee. Raising the SIP 10% a year lifted the return from 20.06% to 20.17% after tax. Step-up SIPs build much more wealth, but that is because they invest much more money (₹805 lakh against ₹257 lakh), not because the rate of return improved.

Run the step-up SIP →

On this index, over this window, almost nothing separated them: lump sum plus SIP returned 20.21% after tax against 20.06% for the pure SIP. The path was different, but the end result was almost the same.

This question separates a real advantage from a lucky start date, so the lab tests every rule across overlapping windows instead of a single run. The standard 20% sell / 35% buy-back / 70% moved rule beat plain SIP in only 12 of 138 ten-year windows (9%), trailing by 1.90% a year on average after tax. Over five-year windows it won 27% of the time.

In other words, it did not keep working: over short and long windows alike, the standard rule mostly cost money.

In fast recoveries. A SIP started in January 2020 made 21.68% plain, but only 17.45% with the drawdown rule. The Covid fall was deep enough to trigger the sell, and the recovery was too fast for a good buy-back. The rule sold near the bottom and missed the rebound while sitting in a liquid fund.

Open the Covid backtest →

A smaller fall. The plain SIP's deepest drawdown was −69.18%. The standard 20/35/70 rule cut that to −49.73%, and the most defensive rule in the sweep to −23.77%. If you are asking about market timing because you are afraid of losing two-thirds of your portfolio's value, that protection, not extra return, is what a rule can reliably give you.