What it does
target value = portfolio total × target weight
trade = target value − current value
drift = current weight − target weight
Hold $42,000, $31,000 and $12,000 against targets of 60/30/10 and the total is $85,000. Position 1 should be $51,000, so it needs $9,000 bought; position 2 should be $25,500, so $5,500 sold; position 3 should be $8,500, so $3,500 sold.
The buys and sells always net to zero, because rebalancing moves money between holdings rather than adding it.
Drift is the number that decides whether to act
The trade amounts tell you what to do. The drift column tells you whether it is worth doing.
Rebalancing has costs — spreads, commissions and, in a taxable account, realised gains. A position 0.4 percentage points off target is not worth paying any of those. One 8 points off is a different matter.
Two common rules, both defensible:
- Threshold: rebalance a position when it drifts more than a set amount — 5 percentage points, or a quarter of its own target weight.
- Calendar: rebalance on a fixed date, once or twice a year, whatever the drift.
The evidence for either being superior is thin. What is clear is that doing it constantly costs more than it returns, and never doing it lets one winner quietly become the whole portfolio.
Why it feels wrong, and why that is the mechanism
Rebalancing sells the position that has done best and buys the one that has done worst. It will feel like a mistake almost every time, and it is precisely the behaviour being paid for: it enforces selling high and buying low on a schedule, removing the judgement that usually gets this backwards.
It is also a risk control rather than a return strategy. Left alone, the best performer grows until the portfolio's outcome depends on it alone — which may be the position you understand least, having bought it when it was small.
The tax constraint, which changes the method
In a taxable account every sale realises a gain, and the tax is paid now rather than later. That cost is often larger than the benefit of a small correction.
Two ways round it that cost nothing:
- Rebalance with new money. Direct contributions to the underweight positions and let the drift close without selling anything.
- Rebalance with the income. Dividends and interest can be directed the same way instead of being reinvested where they came from.
Both work well while the portfolio is still growing and stop working once contributions are small relative to the total. Inside a tax-sheltered account none of this applies and drift thresholds can be tighter.