Rebalancing on bands rather than on a date

Rebalancing is usually described as an annual task. An alternative is to act when weights drift beyond a stated tolerance, which changes when trades happen and how many there are.

The two rules answer different questions

A calendar rule asks whether it is time. A band rule asks whether the portfolio has actually moved far enough from target to be worth correcting.

Calendar rebalancing can therefore trade when nothing has drifted, incurring costs for a negligible correction. It can also leave a large drift untouched for months.

Band rebalancing acts only when the deviation is material, which links the trade to the condition it is meant to address rather than to the date.

Choosing a band width

The band is a tolerance around each target weight, expressed either in absolute terms or relative to the weight itself. Wider bands mean fewer trades and larger drifts tolerated.

Relative bands scale sensibly across a portfolio with very different weightings. A fixed absolute band applied to a small holding may never be breached at all.

There is no correct width. The choice trades trading cost against how far the portfolio is allowed to depart from the allocation that was actually decided on.

Bands require monitoring, calendars do not

The practical cost of a band rule is that weights must be checked to know whether a band has been breached. A calendar rule needs no monitoring at all.

Frequent checking is exactly the behaviour a long-horizon investor is usually trying to reduce, and a band rule supplies a legitimate-sounding reason to keep looking.

A hybrid addresses this by checking on a schedule and trading only if a band is breached. Monitoring is bounded and trades still depend on the condition.

Bands concentrate trading in volatile periods

Breaches cluster when markets move sharply, which means band rebalancing prompts action during exactly the episodes that are hardest to act calmly in.

The trade required is also the uncomfortable direction: buying what has fallen and selling what has risen. That is the mechanism working, and it does not feel like it.

Deciding the band in advance and writing it down is what makes the trade executable. Deciding what counts as a breach during the episode is not a rule at all.

Contributions do the work more cheaply

For a portfolio still receiving contributions, directing new money to whichever holding is furthest below target corrects drift without selling anything.

That approach avoids disposals entirely, which matters where a sale has consequences that depend on local rules. Those rules vary by jurisdiction and change.

Its limitation is capacity. Once contributions are small relative to the portfolio, they can no longer correct a substantial drift, and an explicit rule becomes necessary again.