Rebalancing Without Emotion: A System That Survived Two Cycles

Rebalancing means returning a position to its target share after price movement has changed it. Mechanically it is subtraction. Behaviourally it is the hardest thing in this entire practice, because it requires selling what has gone up and buying what has gone down, at exactly the moments when both feel wrong.

Here is the system I have run since 2020, with the failure modes I hit first.

The rules

Target. Each asset has a target share of investable assets, written down, set in a calm month.

Band. I do nothing while the position stays within a set distance of target. Rebalancing on every small drift generates fees and taxable events for no benefit.

Date. The check happens on the first working day of each quarter. Not when the price moves. Not when I think about it.

Direction. If above the band, sell back to target. If below, buy back to target, but only from designated cash, never by selling something else I wanted to keep.

Documentation. Every action gets a line: date, asset, amount, price, reason. The reason is always “quarterly rebalance”, which is the point.

The failure modes I hit

Skipping a rebalance because the trend looked strong. Did this in late 2021, reasoning that trimming into strength was leaving money on the table. The position was well above band. Three months later it was well below target and I had neither the proceeds nor the discipline.

This is the classic failure. The strongest argument against rebalancing always arrives precisely when rebalancing matters most.

Rebalancing early because a decline looked like an opportunity. Did this in 2022, six weeks ahead of schedule. It was not wrong in outcome. It was wrong in process, and the process is what I am actually maintaining.

Using the wrong denominator. For a period I calculated shares against my crypto holdings only, rather than against total investable assets. That made the whole exercise meaningless, because a position can hold a stable share of a portfolio that has itself become far too large.

Ignoring tax. Rebalancing creates disposals. In most jurisdictions a sale is a taxable event whether or not proceeds leave the asset class. I now consider the tax position before the trade rather than in April.

Why the band matters more than the target

The target is arbitrary within a reasonable range. Twenty percent and twenty-five percent are not meaningfully different decisions.

The band is what determines how often you act, and acting is where the costs and the errors live. A wide band means fewer transactions, less tax and fewer opportunities to talk yourself into something. A narrow band means constant small activity and constant exposure to your own judgement.

Mine is deliberately wide.

The execution detail

A rebalance that requires retrieving a hardware device, remembering a passphrase and waiting for confirmations will get deferred. A deferred rebalance is a skipped one.

So the working balance lives on somewhere with a withdrawal process that works, sized to cover a typical quarterly adjustment, and the cold storage only gets touched when the adjustment is unusually large.

Friction is useful for stopping impulsive trades and harmful for stopping scheduled ones. Design for both.

What it has actually produced

Slightly lower returns than never rebalancing, in the specific period I have run it, because the assets went up and trimming into a rise costs you.

It also produced a portfolio I held through two drawdowns without a single unplanned decision, and the reason I still hold anything at all is that I was never in a position I had not chosen.

Those two facts are the same fact.