Rebalancing does not make you richer; it keeps you holding the amount of risk you chose.

Rebalancing means putting your mix of investments back to the proportions you chose. If you chose six parts shares to four parts bonds, and a good run for shares has carried that to seven parts in ten, rebalancing sells some shares and buys bonds until you are back at six in ten. Nobody has to decide anything for the mix to drift: it happens because the two parts grow at different speeds.

What happens if you never rebalance

Vanguard’s study of a 60/40 US portfolio from 1926 to 2009 is the clearest picture. Rebalanced every year, it averaged 60.5% in shares, returned 8.6% a year and had a volatility of 11.9%. Never rebalanced, it drifted to an average of 84.1% in shares — at one point about 99% — and returned 9.1% with a volatility of 14.4%¹. The untouched portfolio earned more, because shares earned more. But it was no longer a 60/40 portfolio: it had quietly become an all-share one, with the falls to match.

That is the honest case for rebalancing. It is not a way to earn more; over long periods in which shares beat bonds, it usually earns a little less. It is a way to keep the portfolio’s risk at the level you decided you could live with — which matters most just before and after you start drawing on it, when a fall that a 60/40 portfolio would shrug off can do lasting damage to one that has drifted to 80/20.

A worked 60/40: 2021 to 2025

Take an illustrative 100,000 — in any currency — split 60/40 at the start of 2021 between the S&P 500 and the Bloomberg US Aggregate bond index, using each index’s calendar-year return^2,3. One version is never touched. The other is put back to 60/40 every January.

By the end of 2025 the untouched portfolio was worth 156,981 and held 75.0% in shares; the rebalanced one was worth 151,591^2,3. Rebalancing cost 5,390 over five bull-market years — the price of selling shares that kept rising. Now suppose shares fall 25%, the size of the dot-com style slump in Nivritee’s stress tests⁴. Back at 60/40 in January, the rebalanced portfolio would lose 22,739, or 15.0%. The untouched one, at 75/25, would lose 29,425, or 18.7% — 6,687 more, which is more than the 5,390 it had gained by not rebalancing, given back in one bad year. Morningstar’s Amy Arnott found the same drift in real time: a 60/40 test portfolio left alone for the five years to November 2024 had grown to 73.4% in equities⁵.

How often: calendar, threshold, or both

There are two ways to decide when, and they combine well.

• By the calendar. Look once a year — a birthday, the start of the tax year — and rebalance whatever you find. Vanguard found no optimal frequency, and that annual rebalancing is likely to be preferred where taxes or time are involved¹. William Bernstein, testing a four-asset portfolio over 24 overlapping 28-year periods, found the differences between monthly, quarterly, yearly and longer schedules small, and monthly too frequent⁶.
• By a band. Act only when an asset class has drifted by more than a set amount. Vanguard’s 2010 conclusion was that annual or semi-annual monitoring, rebalancing at 5% thresholds, strikes a reasonable balance between risk control and cost¹. Its 2024 paper on target-date funds found threshold rebalancing added 15–22 basis points a year over monthly rebalancing while saving, and 22–25 while drawing down⁷.
• The 5/25 rule. Larry Swedroe’s version, as Ben Carlson summarises it: rebalance when an asset class moves more than 5 percentage points from its target, or 25% of its target, whichever is smaller. A 60% share holding is rebalanced outside 55–65%; a 10% holding outside 7.5–12.5%⁸.

Where you rebalance to matters as well as when. When the cost is your time, going all the way back to the target makes sense; when the cost is proportional — commissions, taxes — Vanguard found it better to trade only back to the edge of the band¹. Its 2024 paper uses the same idea for its target-date funds: trade when an allocation drifts 2 percentage points from target, and move it back to 1.75 points away rather than all the way — so a 60/40 portfolio breaching 62% in shares would go back to 61.75%, not 60%, which keeps each trade small⁷.

Is there a rebalancing bonus?

Sometimes, and it is smaller than it sounds. Bernstein found that from 1926 to 1994, a 50/50 mix of shares and long-term corporate bonds rebalanced every year returned 8.34%, against 7.85% for the simple average of the two assets’ returns — a “bonus” of 0.49% a year⁹. The bonus is real when two assets are both volatile and poorly correlated: selling what has risen and buying what has fallen harvests their swings. Willenbrock showed that this diversification return comes from the rebalancing itself, not from lower variance¹⁰.

The catch is in Bernstein’s own figures. The same 50/50 portfolio never rebalanced returned 9.17% — more than the rebalanced one — because it drifted into shares, which earned more⁹. The bonus is measured against an average, not against letting it ride. Between two assets of very different expected returns, such as shares and bonds, expect rebalancing to cost a little return and buy a lot of steadiness. Between similar assets — markets in two countries, say — the bonus is more likely to be a true gain.

Rebalance with new money first, and mind the tax

The cheapest rebalancing involves no selling. Vanguard’s 1926–2009 simulation also tested a portfolio rebalanced only by sending its dividends and interest to whichever asset was underweight: it averaged 61.0% in shares, with a volatility of 11.3%, without a single sale¹. The same works with monthly contributions while saving, and with withdrawals once you are drawing down — take them from whatever is overweight first, as Vanguard’s own guidance suggests¹¹. A regular investment plan that is pointed at the lagging asset each month is a rebalancing rule that never triggers a tax bill.

When you must sell, where you sell matters more than when:

• India. Since the July 2024 Budget, long-term gains on listed equity and equity funds held over a year are taxed at 12.5% above ₹1.25 lakh a year, and short-term gains at 20%¹². Switching between two mutual fund schemes, or from a regular to a direct plan, is treated as a redemption and a purchase, so it can create a taxable gain¹³. Using the yearly exemption to rebalance a little each year is one common approach.
• United Kingdom. Nothing in an ISA is subject to capital gains tax, and nor are gilts¹⁴; outside a shelter, gains above the £3,000 annual allowance are taxed¹⁵. Rebalance inside the ISA or pension first.
• United States and elsewhere. The same principle: Vanguard’s study notes that taxes come into it when rebalancing within taxable accounts¹, so do the selling inside retirement accounts where you can and use new money in the taxable ones.

Glide paths: rebalancing toward a moving target

A target-date fund rebalances to a target that changes with age. Vanguard’s glide path holds 90% in shares at 20, begins reducing in midcareer around 40, reaches 50% at 65 and settles at 30% shares and 70% bonds by 72¹⁶. You can follow the same shape yourself: rebalance as usual, but move the target a few points toward bonds every few years as independence approaches. See asset allocation by age for how far apart the common rules of thumb are.

A routine that takes an hour a year

1. Write the target down — what share goes to shares, to bonds and deposits, and to anything else such as gold — and a band either side, such as the 5% thresholds Vanguard found reasonable¹.
2. Point every contribution at whatever is furthest below target, every month. Most years this alone keeps you inside the band.
3. Check once a year on a fixed date. If nothing is outside its band, do nothing.
4. If something is outside, trade back toward the target — inside tax shelters first, and in India using the yearly gains exemption where you can.
5. Once you are drawing down, take withdrawals from what is overweight, and keep a runway in cash and bonds so you never sell shares after a fall to pay the bills (sequence of returns risk).