A mixture drifts, and rebalancing is the rule that pulls it back

Proportions move on their own

Suppose a mixture is set up with two components in some deliberate ratio. Time passes and one component grows faster than the other. The ratio is no longer what it was, and nobody chose the new one. It is a residue of past performance rather than a decision, and it will keep moving in whichever direction the faster component happens to run.

An illustration with round numbers. A mixture is set at 60 to 40. Over several years the first component doubles while the second is flat. It is now 120 against 40, which is 75 to 25. The household believes it holds a moderate mixture and holds a considerably more concentrated one.

The drift is always towards whatever has risen most, which means a portfolio left alone becomes progressively more exposed to whatever has recently done well. That is close to the opposite of what most people would choose deliberately. It also tends to happen slowly enough that nobody notices the shape changing until a period arrives in which the newly dominant component behaves badly.

Rebalancing is a rule for reversing the drift

The correction is mechanical. At intervals, or when the proportions move beyond a stated tolerance, some of the component that has grown is sold and the proceeds are moved to the one that has not, restoring the original ratio. The rule contains no forecast and requires no judgement about what will happen next.

What it feels like is uncomfortable, because it always involves reducing the holding that has been doing well and adding to the one that has not. Any instinct about momentum objects. The rule exists precisely to override that instinct with an arithmetic operation, which is most of its value.

It is worth being clear about what the rule is for. It is a risk control, keeping the mixture at the level of exposure that was deliberately chosen. Whether it improves returns is a genuinely contested question and depends heavily on the period examined.

The frequency involves a real trade-off

Rebalancing very often keeps the proportions tight and generates more transactions, each of which may carry costs and, depending on the arrangement and the country, tax consequences. Rebalancing rarely lets the mixture drift further from its intended shape between corrections.

Two common approaches exist. One is calendar-based: check at a fixed interval regardless of what has happened. The other is threshold-based: act only when a component moves beyond some stated distance from its target. Both are defensible and both mainly serve to remove the decision from the moment.

What is not defensible is rebalancing whenever it feels right, since that reintroduces exactly the judgement the rule was designed to replace, and the moments when it feels right are correlated with the moments when it is hardest to act.

Neither version has a strong claim to superiority over the other, and the difference between checking twice a year and checking once is small compared with the difference between having a rule and not having one. Choosing whichever version you will actually follow is a more useful criterion than optimising the interval.

New money can do the work without selling

For a household still adding money regularly, there is a gentler mechanism. Directing new contributions towards whichever component has fallen behind moves the proportions back without selling anything, which avoids transaction costs and any consequences of realising a gain.

This works well while contributions are large relative to the total and progressively less well as the balance grows, since a fixed monthly amount eventually becomes too small to move a large mixture. At that point the drift has to be corrected directly or accepted.

The reverse applies to a household drawing money out, where taking withdrawals from whatever has grown beyond its target achieves the same correction in the other direction.

What rebalancing is not

It is not a way of buying low and selling high in any useful sense, and framing it that way sets up an expectation it cannot meet. In a sustained one-way period it will reduce a winning holding repeatedly and look foolish for years.

It also does nothing about whether the original proportions were sensible. A rule that restores an inappropriate mixture maintains an inappropriate mixture very reliably. The proportions themselves come from the horizon and the purpose of the money, which is a prior question.

Costs, tax treatment and available arrangements differ enormously between countries and change over time, and whether any of this applies to a particular household depends on the whole position. This explains the mechanism only; a decision of size belongs with a regulated adviser.