When to Rebalance Your Portfolio

Left alone, a portfolio does not stay where you put it. Whatever performs best grows as a share of the total, and the split you chose quietly becomes a different one. The decision to rebalance is the correction to that drift.

It is widely described as a way to improve returns. That is mostly wrong, and starting from the wrong reason leads to doing it too often, at the wrong times, and with an avoidable tax bill.

What rebalancing is actually for

Rebalancing is risk control, not a return enhancer.

Consider a portfolio set at 70% equities. After a long bull run it might sit at 85% without a single decision being made. Nothing was chosen — the balance simply grew into a more aggressive position, and it did so at the point where valuations were highest and the case for caution strongest.

Selling back to 70% restores the risk level you decided you could tolerate. That is the whole purpose. Whether it also adds return depends on how markets happen to behave, and the honest answer is that it sometimes helps and sometimes costs.

There is a real secondary benefit: it enforces a discipline that is hard to apply by judgement, since it mechanically requires selling what has done well and buying what has done badly. Almost nobody does that voluntarily. But treat it as a bonus rather than the reason.

The portfolio rebalancing calculator works out which trades restore the target, and the asset allocation guide covers how the target should be set in the first place.

Thresholds beat the calendar

Two triggers are in common use, and one is better.

Calendar rebalancing happens on a fixed schedule — annually, say. It is simple and takes no monitoring, but it responds to the date rather than to anything real. It will rebalance a portfolio that barely moved and ignore one that moved sharply three weeks after the review.

Threshold rebalancing triggers when a category drifts more than a set amount from target. A five-percentage-point band around each weight is a common choice. This responds to what actually happened, which is the point.

The practical version is a hybrid: check on a schedule, act only if a band is breached. That gives you the low monitoring burden of the calendar with the responsiveness of the threshold.

Wider bands are better than they sound. Frequent small corrections generate costs and tax for very little risk reduction, and there is no evidence that precision here pays. An annual check with a five-point band is a defensible policy for almost any portfolio.

Tax is the main cost, and it is avoidable

In a taxable account, selling an appreciated holding realises a gain. Rebalancing therefore has a real price, and the mistake is paying it unnecessarily when several routes avoid it.

Rebalance inside sheltered accounts first. Trades within tax-deferred and tax-free accounts create no taxable event at all. If your allocation spans several accounts, you can often restore the overall split entirely with trades inside the sheltered ones and never sell anything taxable.

Direct new money at the underweight. New contributions, dividends and interest can be pointed at whatever is below target. For anyone still saving, this alone handles most drift without a single sale.

Use withdrawals in the other direction. If you are drawing down, take the money from whatever is overweight. The withdrawal you had to make anyway does the rebalancing for free — withdrawal sequencing covers how that interacts with which account to draw from.

Pair with losses. Realised losses offset realised gains. If something is down, selling it to offset a gain elsewhere lets you rebalance at no net tax cost.

Check the holding period. A gain held just under a year is taxed as ordinary income. Waiting a few weeks to cross into long-term treatment can materially change the bill — the cost basis calculator tracks what a sale would actually realise.

When not to rebalance

There are situations where the correct action is nothing.

The drift is small. Inside your bands, leave it. Rebalancing to the exact target achieves nothing worth the friction.

Doing it would trigger a large gain and you have no offset. The tax is certain and immediate; the risk reduction is neither. Sometimes waiting for new contributions to fix it is correct.

Your target changed. If your circumstances shifted, do not rebalance to the old target. Reset the target first, then move toward it.

You would be crossing an income threshold. Realised gains raise taxable income, which can trip Medicare surcharges, push other gains into a higher band, or reduce income-tested benefits. Check the Medicare IRMAA calculator before a large realisation in retirement.

The behavioural part is the hard part

Everything above is arithmetic. The difficulty is that rebalancing asks you to do the opposite of what feels right, at exactly the moment it feels most wrong.

After a strong run, rebalancing means selling the thing that has been working and buying the thing that has not. After a crash, it means buying more of what just fell — during the period when it is most tempting to stop.

That second case is where the strategy earns its reputation, and where it is most often abandoned. Rebalancing into a decline is uncomfortable and is precisely when the risk control matters.

The defence is to decide the policy in advance and write it down: the target weights, the band width, the review date, the account order for trades. A rule set calmly is easier to follow than a decision made in a falling market.

One last caution. Rebalancing assumes the categories are worth returning to — it is not a repair for a bad allocation. If the portfolio is concentrated in a single employer's stock or a single sector, that is a diversification problem, and buying more of a declining concentrated position because a band says so is not risk control. Fix the allocation, then let the policy run.