Calendar or threshold — when do professionals rebalance?

Knowing that rebalancing controls risk leaves an open question: when. The answer is a written rule chosen in advance, because the alternative — deciding in the moment — reliably produces the worst version of the decision. There are two main families, plus the hybrid that most large institutions actually use.

Calendar rebalancing

Rebalance on a fixed schedule: annually, semi-annually, quarterly. Nothing about market conditions enters the decision.

• Strength: trivially simple, auditable, and impossible to argue with in the moment. It can be delegated or automated entirely.
• Weakness: it is blind to size. A quarter in which nothing moved still generates a review; a quarter in which equities moved 25% waits until the date arrives.

Threshold (band) rebalancing

Rebalance whenever a weight leaves a stated band. Two conventions, and the difference between them matters more than people expect:

• Absolute bands — e.g. ±5 percentage points. A 60% equity target acts outside 55–65%.
• Relative bands — e.g. ±25% of the target weight. A 60% target acts outside 45–75%; a 5% target acts outside 3.75–6.25%.

Notice the asymmetry. A ±5 percentage-point absolute band applied to a 5% satellite sleeve would permit that sleeve to double to 10% before triggering — it can never really trigger on the downside at all. Relative bands scale sensibly with sleeve size, which is why multi-sleeve portfolios usually prefer them.

• Strength: trades happen when they matter and not otherwise. Drift is capped by construction.
• Weakness: requires monitoring, and can cluster many trades into a single volatile month.

The hybrid, and what the evidence supports

Most institutional policies use calendar review with threshold action: check monthly or quarterly, trade only if a band has been breached. This bounds both the monitoring burden and the drift.

On the empirical question of which schedule is best, the honest answer is that studies consistently find the differences between reasonable rules to be small, and consistently find that having a rule beats having none. Annual-with-bands, quarterly-with-bands and monthly-with-bands all land in a similar place before costs; after costs, the more frequent rules give some of it back. The specific rule matters less than that it is written down and followed.

A worked comparison

Take a 60/40 target and an illustrative decade with a strong equity run. Rough counts:

• Annual calendar: 10 rebalancing events. Maximum drift within a year could reach 5–8 percentage points before being corrected.
• ±5pp absolute bands, checked quarterly: perhaps 4–6 events over the decade, clustered in the volatile years. Drift capped at 5 points by construction — but the portfolio may sit at 64.9% for years without a trade.
• Quarterly calendar: 40 events. Tightest tracking and roughly four times the trading. In a taxable account that means gains realised four times as often — not four times as large in total, because each trade is correcting a quarter's drift rather than a year's.

Both of the awkward cases in that middle line are easier to see than to describe — the long stretch parked just inside the edge, and the breakout that runs well past it:

Each is defensible. They trade the same three things against each other: tracking tightness, transaction costs, and tax friction. Which trade-off is right depends on the account, the jurisdiction and the size of the portfolio — not on a rule that applies to everyone.

Try it now

1. Monthly bars keep the arithmetic small, so switch both charts below to Monthly. Track the drifting equity weight of a 60/40 month by month across the five years on screen.

2. Mark every month where the equity weight left a 55–65% band. How many trades would a threshold rule have generated? Compare with the 5 that an annual calendar rule would have produced over the same span.
3. Convert one of those bands to a relative band (±25% of 60% = 45–75%). How many of your trigger events survive? Describe what the wider band bought and what it cost.