Rebalancing: How Often Is Enough?

Annually, or on a five-point band. That is the answer, and the useful part is why more often is not better, because the reasoning tells you what to do in the year the market makes it difficult.

What rebalancing is for

Rebalancing means selling some of what grew and buying what did not, to return the portfolio to the mix you chose. The SEC's definition is exactly that and nothing more.

Notice what it is not. It is not a way to make more money. In any stretch where stocks beat bonds, and that is most long stretches, rebalancing takes money out of the thing that is winning and puts it into the thing that is not. The portfolio ends with less.

What it buys is the allocation you actually chose. Left alone, a portfolio drifts towards whatever has grown fastest, which means it drifts towards more risk, and it does so silently.

After twenty years the investor holds 82% stocks. They never decided that. They decided 60% once, and then twenty years of arithmetic decided the rest, and the decision surfaces the first time markets fall hard, on a portfolio far more exposed than they believe it to be. That is the risk rebalancing controls, and it is worth reading beside how the allocation is chosen in the first place.

Calendar against threshold

Calendar. A fixed date. Simple, automatable, requires no monitoring, and occasionally trades when nothing has happened.

Threshold, or band. Act when a holding is more than a set distance from target, usually five percentage points. Trades only when something moves, and requires you to look. With a three-fund portfolio that is three numbers to read once a year.

The practical version combines them: check once a year, act only if something is outside the band. That is one decision a year and usually the decision is to do nothing.

What frequent rebalancing costs

A $200,000 portfolio at 60/40, stocks returning 9% and bonds 3%, run for twenty years in a taxable account, with realised gains taxed at 15%.

Three readings, in order of importance.

Quarterly is not worth it. It finishes $3,663 behind annual after tax, on $200,000 over twenty years, for four times the decisions. That is a difference of about half a percent of the final balance, earned by doing sixty more things.

The band beat the calendar. Five trades, $14,575 more after tax than annual, and the same 60% allocation at the end. It traded when the portfolio had actually moved rather than when the calendar said so.

Never rebalancing finished with the most money. It has to, in a model where stocks beat bonds every year, and the same is roughly true of the real period. That is not an argument against rebalancing, it is the price of it: about $95,000 over twenty years, paid for holding 60% stocks instead of 82%. If the 82% portfolio would have survived the fall that happens in year fourteen of a real twenty years, it was the better choice. That is not knowable in advance, which is what the 60% was for.

Rebalancing with new contributions instead

The cheapest correction is the one that never sells anything. If you are still paying in, point this month's contribution at whatever is below target.

On a $200,000 portfolio drifting 1.4 points a year, the drift is about $2,800. Contributions of $1,000 a month directed entirely to bonds fix that in three months, at no tax and no transaction cost. For most people still contributing, this handles ordinary drift indefinitely and the annual check becomes a confirmation rather than a trade.

Withdrawals work the same way in reverse: take the income from whatever is above target.

Taxable accounts: the wrinkle

In an IRA or a workplace plan, rebalancing realises nothing and costs nothing in tax. In a taxable account, selling the winner realises a capital gain in that year, at long-term rates if you held it over a year and at ordinary rates if you did not.

Three consequences follow.

• Do the trading in the sheltered account. If your stocks and bonds sit across a 401(k) and a brokerage account, rebalance inside the 401(k) and leave the brokerage alone. The overall allocation is what matters, not each account's.
• Use contributions and withdrawals in the taxable account. Free, and usually sufficient.
• Watch the holding period. Selling a position held eleven months converts a long-term gain into a short-term one taxed as ordinary income. Waiting a month is often worth more than the rebalance.

If you hold a target date fund, none of this applies to you: the fund rebalances internally and no gain is realised in your account when it does.

When this does not apply

If your allocation should change. Rebalancing returns you to a target. If the target is wrong for your circumstances, fix the target first and then rebalance to it.

If the portfolio is small. Under roughly $10,000, a five-point drift is $500 and the whole exercise is noise. Direct contributions and get on with your life.

If you hold one fund. A target date fund or a balanced fund already does it. Rebalancing between two such funds is a decision about nothing, and the same goes for swapping between an index fund and its ETF twin.

If a large gain is concentrated in one taxable position. Selling to rebalance may trigger a tax bill worth more than the risk reduction. Stopping reinvested dividends and directing new money elsewhere gets you there without the event.

If you are about to draw on it. Inside a few years of spending the money, the question is not drift, it is having the next few years of spending in something that cannot fall.

Next steps

1. Write down your target allocation. If you cannot, that is the job for today, not rebalancing.
2. Work out your current allocation across every account together, not one at a time.
3. If anything is more than five points from target, correct it now.
4. Point your next few contributions at whatever is light, before you sell anything.
5. Do the selling inside a sheltered account wherever possible.
6. Put one date a year in the calendar, and on that date check the band rather than trade automatically.