Allocation drift: the risk you didn't choose

A friend set up a tidy 60/40 three years ago. Equity for the growth, debt so he could sleep at night. He was proud of it, and then he never touched it again.

Grown nicely, he told me. I asked him what the split was now. He didn't know. So we opened the statement, and it was not 60/40 anymore.

Here is the quiet mechanism nobody warns you about. Equity compounds faster than debt, so its share of your portfolio creeps up every single year you leave the mix alone. Assume equity earns about 12 a year and debt about 6.5. A 60/40 you never rebalance drifts to roughly 66/34 in five years, and about 71/29 in ten. Eleven points more equity, eleven points more risk, and not one of those points was a decision you made.

Now the part that actually costs you. That equity share is at its highest after a long run up, which is exactly when a correction is most likely. When the market finally turns, say equity down 40 and debt up 2, the drifted 71/29 falls about 27.8 percent. The 60/40 you actually chose falls 23.2. That is nearly 5 extra points of drawdown on a bet you never placed, and it needs a bigger bounce just to climb back to even.

This is why I think rebalancing gets sold the wrong way. Everyone frames it as a free-return trick, buy low and sell high. The bigger job is quieter. It drags your mix back to the risk level you picked on a calm day, before the market talked you into carrying more.

Leaving it alone is not caution. It is handing the rally your risk dial, and on the way up the rally turns it all the way to maximum:

Drift runs both ways: a real 60/40 from Oct 2024

The ten year example above uses assumed returns. So I ran the same question on actual NSE closes, over a window that did not go up in a straight line.

Put Rs 10,00,000 into a 60/40 on 14 Oct 2024: Rs 6,00,000 tracking the Nifty 50 and Rs 4,00,000 tracking the Nifty Composite G-sec Index. The Nifty 50 closed at 25,127.95 that day and the G-sec index at 2,828.72. Then leave it alone.

By 30 Sep 2026 the Nifty 50 closed at 22,620.45, down 9.98 percent. The G-sec index closed at 3,101.54, up 9.64 percent. The equity sleeve is worth about Rs 5,40,126, the debt sleeve about Rs 4,38,579. The mix is now 55.2 percent equity and 44.8 percent debt.

Same mechanism, opposite direction. After a fall the mix holds less equity than you chose, so a rebalance on 30 Sep 2026 would have bought about Rs 47,000 of equity and sold the same amount of debt to restore 60/40. That is the trade almost nobody wants to make after two flat years. It is also the trade that keeps the plan honest. One caveat on the numbers: the Nifty 50 here is the price index, and its dividend yield was 1.22 on 30 Sep 2026 by the NSE data, so a real equity sleeve did a little better than this line.

Weight, drift and the band

You do not need a statement to see drift coming. One formula does it. The equity weight after a period is the equity value grown by its return, divided by the whole portfolio grown by both returns.

For the friend in the opening: 60 percent equity grown at 12 percent for five years is 60 x 1.762 = 105.7. Forty percent debt grown at 6.5 percent is 40 x 1.370 = 54.8. Equity is 105.7 out of 160.5, which is 65.9 percent. That is the 66/34 in the post.

A band rule turns that number into a decision. Pick a band, say 5 points either side of 60. Inside the band you do nothing. Outside it, you trade back to target. On the real NSE window above, the mix sat inside a 5 point band the whole time and touched 55.2 only at the very end, so a band investor would have been right at the edge on 30 Sep 2026. A calendar investor rebalancing every year would have traded twice. The drift threshold calculator sizes the trade for your own numbers.

What rebalancing is for

The free Learn module on rebalancing strategies works through calendar, band and signal driven rules on a 60/40 in rupees, with the tax side of every trade. It also has a quiz and a certificate if you want to test yourself.

The friction worth knowing

Rebalancing a mix is not free. Selling equity held 12 months or less is taxed at 20 percent on the gain, and longer holdings at 12.5 percent above Rs 1.25 lakh of gains a year, for transfers from 23 July 2024. Debt fund units bought on or after 1 April 2023 are taxed at your slab rate whatever the holding period.

So the cheapest rebalance is often the one done with new money. An investor who adds Rs 20,000 a month can point it at whichever sleeve is under target, and the mix drifts back without a single sale. Only when cash flow cannot close the gap does a sale earn its tax.

If you want the version of this on a stock basket, rebalance cadence shows how one runner takes over an equal weight book between rebalances.