Markets
Rebalancing Bands: When to Actually Do It, and What It Costs in a Taxable Account
Rebalancing gets sold as a way to "buy low and sell high automatically." That framing is mostly wrong, and believing it leads people to rebalance too often, in the wrong account, and at a tax cost that quietly exceeds whatever benefit they imagined they were capturing. What rebalancing actually does is more modest, more reliable, and worth understanding precisely.
What rebalancing is actually for
Pick a target allocation — say 70% stocks, 30% bonds — and time will destroy it. Stocks outgrow bonds over most multi-year stretches, so the same portfolio drifts to 78/22, then 84/16, without you doing anything.
The portfolio has become riskier than the one you chose. Nobody decided this. It happened because the faster-growing asset became a larger share of the total by arithmetic. Rebalancing sells enough of what grew and buys enough of what lagged to restore the mix you intended.
That is a risk control function, not a return-enhancement function. In fact, over long periods where stocks outperform bonds, rebalancing typically lowers total return slightly, because you are repeatedly trimming the better-performing asset. You accept a small expected return cost in exchange for keeping the portfolio's risk where you can live with it.
There is a real behavioral benefit alongside the mechanical one. Rebalancing forces you to buy the asset that just performed badly, at the moment it feels worst to do so. A rule that makes you buy stocks in March 2020 is valuable precisely because almost nobody does that voluntarily.
Calendar rebalancing versus bands
Two approaches dominate, and one is clearly better reasoned.
Calendar rebalancing means checking on a schedule — annually, semi-annually, quarterly — and restoring targets regardless of how far things have drifted. It is simple and requires no monitoring. Its flaw is that it is arbitrary: your portfolio has no idea it is December. You may trade when nothing has meaningfully moved, incurring costs for nothing, and sit through a large drift for eleven months because the calendar has not come around.
Threshold or band rebalancing means acting when an allocation moves more than a set distance from target, whenever that happens. This ties the action to the condition that actually matters — drift — rather than to the date.
The most widely used formulation is the 5/25 rule. Rebalance an asset class when it deviates by either:
- 5 absolute percentage points from its target, or
- 25% of its target weight, relative
The two triggers handle different position sizes sensibly. A 40% target uses the absolute test, since 5 points arrives before 25% relative (which would be 10 points). A 5% target uses the relative test, since 25% relative is 1.25 points and 5 absolute points would require the position to triple. Whichever triggers first is your signal.
In practice, most people run a hybrid: check on a schedule, act only if a band has been breached. You get the discipline of a calendar without the arbitrary trading.
Rebalancing is a risk-control mechanism that costs a little return. Treating it as a return strategy leads to doing it far too often.
The tax cost that changes everything
Everything above assumes trading is free. In a taxable brokerage account it is not, and the cost can dwarf the benefit.
Selling appreciated stock to buy bonds realizes a capital gain. Long-term gains are taxed at 0%, 15% or 20% depending on taxable income — for 2026, the 15% rate begins above $49,450 of taxable income for single filers and $98,900 for married filing jointly. Short-term gains on anything held a year or less are taxed as ordinary income, up to 37%.
Consider a $400,000 taxable account that has drifted from 70/30 to 78/22. Restoring the target means selling roughly $32,000 of stock. If that stock carries a 50% embedded gain, you realize about $16,000 of long-term gain. At 15%, that is $2,400 in federal tax, plus state income tax, paid immediately and permanently.
What did you buy for $2,400? A modest reduction in portfolio risk. Sometimes that is worth it. Often it is not, and the drift would have been better tolerated or corrected another way.
Rebalance without selling anything
Three techniques avoid the tax cost almost entirely, and they should be exhausted before you consider selling.
Direct new contributions to the underweight asset. If you are still adding money — a monthly investment, a bonus, an annual contribution — send all of it to whatever is below target. For anyone in the accumulation phase with meaningful ongoing contributions, this alone handles most drift and generates zero tax.
Stop automatic dividend reinvestment. Instead of reinvesting each fund's dividends back into itself, let distributions accumulate as cash and deploy them where they are needed. This is a small, steady, tax-free rebalancing current.
Do all selling inside tax-advantaged accounts. This is the most important one. Trades inside a 401(k), IRA, Roth or HSA generate no taxable event whatsoever. You can rebalance as often as you like there, at no tax cost.
The key mental shift is to treat all your accounts as one portfolio rather than several. If your overall stock allocation is too high, you do not need to sell stock in the taxable account — you can sell stock and buy bonds inside the IRA, and the household allocation is corrected without a single taxable dollar.
Asset location makes this easier
Which naturally leads to asset location: deciding which assets live in which account type. Tax-inefficient holdings — taxable bonds, REITs, high-turnover active funds — generally belong in tax-advantaged accounts, where their income is shielded. Tax-efficient holdings such as broad stock index funds are reasonable in taxable accounts, since they distribute little and qualify for long-term rates.
A useful side effect is that keeping bonds in the IRA and stocks in the taxable account puts the asset you will most often need to buy when rebalancing right where trading is free. The structure does the work for you.
One caution: do not let tax location override the allocation itself. Ending up with an all-stock taxable account and an all-bond IRA is fine only if the combined result is the mix you actually wanted.
A policy you can write down
Decide these four things once, write them down, and follow them rather than re-deciding under market stress:
- Target allocation, with a specific percentage for each asset class.
- Trigger: 5 absolute points or 25% relative, whichever comes first.
- Review cadence: a specific date, once or twice a year, to check whether a band was breached. Not a date to trade.
- Order of operations: new contributions first, then dividends, then trades inside tax-advantaged accounts, and only then consider realizing gains in taxable — and if so, harvest offsetting losses in the same year where possible.
Most years, following this policy means doing nothing at all. That is the correct outcome, not a failure to manage the portfolio.
Key takeaways
- Rebalancing controls risk; it usually reduces long-run return slightly rather than adding to it.
- Threshold bands beat calendar dates because they respond to actual drift — the 5/25 rule covers both large and small positions.
- In a taxable account, restoring a target can trigger thousands in capital gains tax, which frequently exceeds the benefit.
- New contributions, redirected dividends and trades inside tax-advantaged accounts rebalance at zero tax cost — use all three first.
- Treat every account as one portfolio: correcting the household allocation inside an IRA avoids selling in taxable entirely.
- Write the policy down in advance so the decision is made before the market makes it emotional.