How to Rebalance a Portfolio
To rebalance a portfolio, compare each holding's current weight against your written target allocation, then sell the overweight positions and buy the underweight ones until the weights match again. Do this on a fixed rule, a date or a drift band, rather than when it feels necessary.
Rebalancing means returning a portfolio to weights you decided earlier. Mechanically it is subtraction. The difficulty is entirely psychological: it requires selling whatever has worked and buying whatever has not, on a schedule, without an opinion.
Before you start
A written target allocation, because rebalancing needs something to rebalance toward. Percentages per asset class, written down, ideally somewhere you cannot casually edit.
A rule for when you do it, decided before any drift exists. Either a date, or a band, or both. The rule has to predate the drift or it will be written to accommodate it.
The tax position of the account, since selling in a taxable one has a cost the sheltered version does not. This changes how wide the bands should be, not whether you rebalance.
The steps
1. Write the target weights down
Sixty and forty, or whatever the allocation is. Without a written number, “rebalancing” becomes adjusting toward whatever seems reasonable today, which is not the same activity.
2. Choose the trigger before you need it
Annually on a fixed date is the simplest. A band — act when any holding drifts more than five percentage points from target — responds to markets rather than to the calendar. Both work; choosing after the drift does not.
3. Measure the current weights
Current value of each holding divided by the total. This is the only calculation involved, and doing it is often enough to reveal drift nobody had noticed.
4. Use new contributions before selling anything
Direct the next contribution into the underweight holdings. In a taxable account this is materially cheaper than selling, because it realises nothing.
5. Sell the overweight only if buying cannot close the gap
Contributions cannot fix a large drift in a big portfolio. When selling is required, sell from the sheltered account first if you hold the same assets in both.
6. Rebalance across accounts, not inside each one
Your allocation is a fact about everything you own. Balancing each account separately forces sales in taxable accounts that a whole-portfolio view would have avoided.
7. Record the date and the weights afterwards
Two lines in a file. It makes the next rebalance a comparison rather than a fresh calculation, and it is the evidence that the rule is actually being followed.
How to tell it worked
A written target exists and was not edited during the process.
The trigger rule was decided before the drift, on a date or a band.
Weights are within your band of target, measured after the trades settled.
And the whole thing took under 30 minutes, because it is arithmetic rather than a judgement.
What it costs
Transactions and, in a taxable account, realised gains. On this site’s arithmetic a 75-basis-point annual drag removes 20.2% of a thirty-year pot, and over-frequent rebalancing is one of the ways a portfolio acquires a drag it did not intend.
Which is why annual is usually enough. Quarterly costs three times as much in transactions for a difference in weights that is generally small.
Why it feels wrong every time
You are selling the thing that has gone up. That is the mechanism, not a side effect: the overweight holding is overweight because it performed, and reducing it is the whole point.
Every argument against doing it is available and sounds reasonable. Momentum, the position is working, the other asset is falling for a reason. All of these may be true and none of them is a reason to abandon a target you set deliberately.
The rule exists because the judgement is unreliable at exactly this moment. A date on a calendar does not know what has been working, which is its entire advantage over you.
Bands against dates
A calendar rule is the simplest thing that works. One date a year, the same date every year, no judgement involved. Its weakness is that it can act when nothing has drifted and sit still through a year when something drifted a great deal.
A band rule responds to the portfolio instead. Act when any holding is more than five percentage points from its target, whenever that happens. It costs nothing in a quiet year and reacts in a volatile one.
Both together is the usual compromise. Check on a fixed date, act only if something is outside its band. That is one diary entry a year and a rule that does nothing when nothing needs doing.
What matters far more than the choice is that it was made in advance. Any of the three beats deciding in the moment, because the moment always argues for leaving the winner alone.
The original data
Of the 24,971 unique videos in research/search-study-corpus.jsonl, 1 mentions rebalancing in the
title, at 264 views — and it is not instruction-shaped. Portfolios generally appear in 147 at a median
of 11,537 and asset allocation in 4 at 5,289. The counts come from site/corpus_count.py and
site/rank_investing.py.
1 video at 264 views, against 147 on portfolios at 11,537. The maintenance procedure that keeps an allocation being the allocation has essentially no coverage, while the decision about what the allocation should be has plenty.
The answer to the question on that chart is that the band was set for exactly this situation. Twelve points over is drift, and it is drift caused by the holding doing well, which is the only way drift happens. Skipping the trim because it is still rising is deciding the allocation by recent performance, which is what the target existed to prevent.
When it fails
The failure is the postponed rebalance, and it takes years to become visible. The date arrives, one holding is well above target and performing, and trimming it feels obviously wrong. It gets postponed until next quarter, then the year after. Five years on, the portfolio is concentrated in whatever ran hardest, the allocation on paper bears no relation to what is held, and the risk profile is one nobody chose — it was assembled by momentum, one deferral at a time.
The second failure is no written target. There is nothing to rebalance to.
A third is choosing the trigger after seeing the drift. The rule then accommodates it.
A fourth is balancing each account separately. It forces avoidable taxable sales.
A fifth is rebalancing quarterly by default. Three times the cost, marginal benefit.
And a sixth is selling before using new contributions. The cheaper tool goes first.
Related
Asset allocation is where the target weights come from. Diversification is what the target is trying to preserve. And portfolio building covers assembling the holdings in the first place.
It has never once felt like a good idea at the time. I am always selling the thing that has done well to buy the thing that has not, and every argument for skipping it is available and persuasive. That is exactly why it has to be a rule with a date on it rather than a decision I make while looking at the numbers.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.