Rebalancing instructs an investor to sell part of whatever has risen most and buy more of whatever has lagged. The discomfort this produces is the reason it is usually done badly or not at all.
Portfolios drift away from their targets
A portfolio set at fixed proportions does not stay there. Assets that rise become a larger share of the total simply by rising.
After a strong period in one asset class, an allocation chosen for a particular risk level can hold considerably more of that asset than intended.
Nothing signals this. The portfolio still contains the same holdings, and the change is in the proportions rather than the contents.
Drift raises risk exactly when it feels lowest
The drift always runs towards whatever has performed well, which is also whatever feels safest at that moment because the recent experience of holding it has been good.
The portfolio therefore carries its largest exposure to an asset after that asset has already risen, which is when the exposure is least likely to be rewarded.
Rebalancing corrects this mechanically, without requiring any view about what will happen next, which is the feature that makes it usable.
Why it is not a return strategy
Rebalancing is often described as buying low and selling high, which overstates it. Its reliable effect is on risk control rather than on returns.
In extended trends, rebalancing reduces returns because it repeatedly trims the asset that continues to rise. In choppier markets it can add a small amount.
The honest case for it is that it keeps the portfolio matched to the risk the investor decided they could tolerate, which is a different objective from maximising the outcome.
How the trigger is chosen
Calendar rebalancing acts at fixed intervals regardless of conditions, which is simple and can act when nothing needs doing.
Threshold rebalancing acts when an allocation drifts beyond a set band, which responds to actual movement and requires monitoring.
Both work. The evidence for a specific interval or band being superior is weak, which suggests the choice matters far less than whether a rule is followed at all.
Where the costs sit
Every rebalance involves trading, so spreads and commissions apply, and in taxable accounts a sale can create a taxable event whose treatment varies by jurisdiction.
Directing new contributions towards the underweight assets achieves much of the same effect without selling anything, which is why it is the cheaper route while a portfolio is still growing.
Frequency therefore involves a trade-off between staying close to the target and paying to get there, and rebalancing too often costs more than the drift it corrects.