Rebalancing is the unglamorous maintenance job that quietly does most of the work in a long-term portfolio. The idea is simple: your target mix drifts as some assets run and others lag, so you periodically sell a bit of what’s grown and buy a bit of what’s fallen to get back to plan. The hard part isn’t the concept — it’s deciding when to do it without turning a risk-control habit into a fee-and-tax machine.

There are two common triggers: a calendar (rebalance on fixed dates) and threshold bands (rebalance only when an allocation drifts past a set distance from target). They sound similar. In practice they behave quite differently, and one of them tends to waste a lot less money.

What rebalancing is actually for

Rebalancing is a risk tool first and a return tool a distant second. Left alone, a 60/40 portfolio in a long bull market slowly becomes 75/25 — and now you’re carrying far more risk than you signed up for, right before you find out the hard way. Rebalancing drags that risk back to target. Any “return benefit” from selling high and buying low is real but small and inconsistent; don’t rebalance for returns, rebalance to keep your risk where you decided it should be.

That framing matters because it changes the answer to “how often?” If the job is risk control, you only need to act when risk has actually drifted — not on a schedule for its own sake.

Calendar rebalancing

Pick an interval — quarterly, semi-annually, annually — and on that date, reset everything to target regardless of how far it drifted.

  • Upside: dead simple, fully automatable, removes judgment and emotion. You never forget, and you never agonize.
  • Downside: it fires whether or not anything meaningfully moved. A calendar rebalance in a flat quarter just generates trades — and therefore fees, spreads, and (in a taxable account) realized gains — for almost no risk benefit. It can also miss a big mid-period swing because the next date is months away.

Annual calendar rebalancing is the most defensible version: cheap, low-touch, and frequent enough to matter. Quarterly is usually overkill for a long-term portfolio.

Threshold-band rebalancing

Set a tolerance around each target — say, “rebalance crypto whenever it drifts more than 5 percentage points from its 15% target” (i.e. outside the 10–20% band). You only act when an allocation actually leaves its band.

  • Upside: it triggers on the thing you care about — real drift — and stays quiet when nothing’s happening. In calm markets it might not fire for a year; in a violent one it fires exactly when risk has moved, even if that’s two weeks after your last check.
  • Downside: you have to monitor (easy to automate, annoying by hand), and very tight bands can trigger constantly in choppy markets, which reintroduces the cost problem.

Wider bands = fewer, cheaper trades and more tolerance for noise. Tighter bands = closer tracking but more friction. For most long-term investors, bands in the range of ±20% relative to the target (or roughly ±5 percentage points on a mid-size sleeve) strike a sane balance.

The hybrid most pros actually use

The cleanest approach combines both: check on a schedule, but only trade when a band is breached. You look monthly or quarterly (or your script does), and you rebalance only the sleeves that have actually drifted out of tolerance. You get the discipline of the calendar and the cost-efficiency of bands, without rebalancing things that haven’t moved.

This is also far gentler on a taxable account, because you’re not realizing gains on a fixed date for no reason — you only sell when risk genuinely requires it.

Cut the cost before you cut the frequency

Before obsessing over the trigger, use the free levers that avoid selling entirely:

  • Rebalance with new contributions. Direct fresh money to whatever’s under target. If you’re still adding monthly, you can often keep the mix in line for years without selling a thing — no taxes, no realized gains. This pairs naturally with a DCA approach.
  • Rebalance in tax-advantaged accounts first. Trades inside an IRA or equivalent don’t trigger taxable events, so do your selling there when you can.
  • Mind the band on the volatile sleeve. Crypto and small experimental positions drift fastest; they’re usually where rebalancing earns its keep. A stable bond sleeve rarely needs touching.

A workable default

If you want one rule to start with: check quarterly, rebalance only sleeves that have drifted past their band, and steer new contributions to underweight positions in between. Use wider bands on volatile assets, set hard caps you won’t override, and write the whole thing down so a calm version of you decides the rules before a stressed version of you has to follow them. That last part is the real point of automating any of this — it’s covered more deeply in the practical automated investing stack and risk-first portfolio automation.

FAQ

How often should I rebalance?

As often as your risk actually drifts — not on a fixed schedule for its own sake. Checking quarterly and trading only when an allocation breaches its band is a sensible default; annual calendar rebalancing is a fine simpler alternative.

Does rebalancing improve returns?

Sometimes, slightly, but inconsistently. Treat any return benefit as a bonus. The reliable benefit is keeping your portfolio’s risk at the level you chose.

What’s a reasonable threshold band?

Roughly ±5 percentage points on a mid-size allocation, or about ±20% relative to the target weight. Wider on volatile sleeves, tighter on stable ones. Avoid very tight bands — they trigger constantly and bleed fees and taxes.

Bottom line

Rebalancing is risk maintenance, not a profit engine, so do it when risk drifts — not because the calendar said so. Threshold bands (ideally checked on a schedule and topped up with new contributions) keep your mix honest while wasting far less on fees and taxes than rigid quarterly resets. Less really is more here.