How Often Should You Rebalance Your Portfolio? What the Evidence Says
Key takeaway
Rebalance about once a year, and only past a five-point drift. Rebalancing keeps risk at target; it does not reliably add return. What the research finds, what drift costs in drawdown, and how to do it without realising a gain.
The short answer: rebalance about once a year, and only if a sleeve has drifted more than five percentage points from its target. The longer answer is that rebalancing is not a return strategy, whatever the “rebalancing bonus” articles say. It is the mechanism that keeps the portfolio you own the same as the portfolio you chose.
That distinction decides everything else. If rebalancing were a way to make money, the right frequency would be whatever captures the most of it. Because it is a way to control risk, the right frequency is whatever holds your risk near target at the lowest cost. Those two questions have different answers, and most guidance quietly answers the first while claiming to answer the second.
Why portfolios drift, and why the drift always goes the same way
Nobody decides to become more aggressive. It happens by not doing anything.
A portfolio drifts toward whatever went up. Because equities go up more often than bonds and by more when they do, the drift is almost always in the direction of more equity risk. Two real periods, with the arithmetic done on a portfolio that started at 60% equities and 40% bonds and was left alone:
| Period | Equities | Bonds | A 60/40 becomes | Drift |
|---|---|---|---|---|
| Calendar 2025 | +14.42% | −0.36% | 63.3 / 36.7 | 3.3 points |
| 12 months to 28 Jan 2026 | +17.88% | +7.30% | 62.2 / 37.8 | 2.2 points |
| Three years at +15% / 0% | +15% a year | flat | 69.5 / 30.5 | 9.5 points |
The first two rows use the S&P 500 and the Bloomberg US Aggregate over the periods named. The third is arithmetic, not a forecast: it shows what an ordinary run of good equity years does when nobody intervenes.
Notice the shape of the problem. A single good year moves the needle three points, which feels like nothing. Three good years in a row move it nearly ten, which is a different portfolio. And the drift is worst precisely at the end of a long expansion, which is when the next decline arrives.
The practical consequence: an unrebalanced portfolio carries its maximum risk at the moment it can least afford to.
What the drift actually costs, in the only unit that matters
Percentage points of allocation are abstract. Convert them into the loss you would sit through.
Global equities fell roughly 55% peak to trough between October 2007 and March 2009. Applying that decline to each equity weight, and assuming the bond sleeve held its value as it broadly did in that episode:
| Equity weight | Loss in a 2008-scale decline | Gain needed to recover |
|---|---|---|
| 60% (your target) | −33% | 49% |
| 65% (after two good years) | −36% | 56% |
| 70% (after three or four) | −38% | 63% |
Ten points of drift is five points of extra drawdown and fourteen extra points of recovery. That is the entire cost of not rebalancing, and it is not paid in returns. It is paid in the size of the hole, which is the thing that decides whether people stay invested at all. Our piece on how emotions affect investment decisions covers what happens next when the hole is deeper than expected.
There is a second cost that is easier to miss. If you took a risk profile assessment and it put you at moderate, a 60/40 that has drifted to 70/30 is no longer a moderate portfolio. You have been moved into the next profile up without deciding anything, and without the assessment that would have told you whether you could hold it.
What the research says about how often
This is where most articles reach for the rebalancing bonus: the claim that systematically selling what rose and buying what fell adds return. Sometimes it does. Over long periods, in the data, it mostly does not, and the effect is dominated by whether the assets mean-revert over the window you happened to measure.
Vanguard’s work on this is the most cited, and its conclusion is narrower than the way it usually gets quoted. Tracking a 60/40 portfolio back to 1926, it found that rebalancing monthly or quarterly produced no meaningful improvement in long-term risk or return compared with rebalancing annually, while generating more transaction costs along the way. Its own guidance settles on an annual rebalance as reasonable for most investors, and its threshold work for target-date funds uses a trigger of two percentage points of drift.
The same research shows what frequency actually buys you, and it is not return. In the March 2020 crash, a 60/40 rebalanced on a monthly calendar could drift as much as seven points from target, and on a quarterly calendar as much as ten, while a threshold approach that fired at two points of drift kept the deviation far smaller. The threshold method is better because it responds to the market rather than to the date, not because it earns more.
The practical consequence: more frequent rebalancing does not pay you. It costs you, in fees and in tax, in exchange for holding risk closer to target. Buy exactly as much of that as you need.
The cost that never appears in the comparison
Every study of rebalancing frequency you will read compares strategies in a tax-free account. Your account is probably not tax-free.
In a taxable account, rebalancing after a good run means selling the thing that rose, which realises a gain and triggers tax now. The benefit of rebalancing is probabilistic and arrives later. The tax is certain and arrives immediately. That asymmetry is why a rule that looks optimal in a study can be actively wrong in your account.
Add the mechanical costs: spreads on whatever you trade, commissions if your broker still charges them, and the currency conversion if the sleeves are denominated differently. None of these is large individually. Applied four times a year to a portfolio that did not need it, they compound against you.
How to rebalance without selling anything
The most useful technique in this whole subject gets one line in most guides.
If you are still contributing, you can rebalance with new money. Direct each contribution to whichever sleeve is furthest below target and let the purchases do the work. No sale, no realised gain, no tax, and the spread you were going to pay anyway on the new money.
For most people in the accumulation phase, contributions are large enough relative to the portfolio that this handles the drift entirely for years. Selling only becomes necessary once the portfolio is large relative to what you add each year, or once you stop adding. The same logic applies to withdrawals in reverse: take income from the sleeve that is above target.
Where you must sell, sell inside a tax-sheltered account first. A pension or its local equivalent can be rebalanced at no tax cost, so if your allocation spans several accounts, use the sheltered one to do the adjusting and leave the taxable one alone.
A rule you can actually follow
The rule has to survive contact with a year in which you do not want to follow it, so it should be short.
- Check once a year, on a date you fix in advance. Your birthday, the first business day of January, whatever you will not negotiate with.
- Act only if a sleeve is more than five points from target. A 60/40 at 63/37 is not a problem. At 66/34 it is.
- Rebalance with contributions where you can, and sell only what the threshold requires. Go back to target, not past it.
- Write down the target and the date. The rule fails at the moment when rebalancing means buying the asset everyone is describing as broken, and a written rule is the only thing that survives that moment.
The five-point threshold is a compromise, not a law. Vanguard’s institutional work fires at two points because a fund can trade at scale for almost nothing. You cannot, so a wider band that trades less often is the right trade for a private investor. If your portfolio is entirely inside a tax-sheltered account and your broker charges nothing, tighten it toward three.
When this does not apply
If you hold a single multi-asset fund or a target-date fund, the manager rebalances inside the fund and there is nothing for you to do. That is most of the argument for owning one.
If your allocation is entirely equities, there is no bond sleeve to drift against and rebalancing means adjusting between regions or factors, where the case is much weaker and the costs are the same. And rebalancing does nothing about position size inside a trading account, which is a different problem with different arithmetic, covered in position sizing and risk of ruin.
A note on the present moment
Equities have run hard into September 2026 while the Federal Reserve holds at 3.50% to 3.75% and long yields sit well above the policy rate, a divergence we took apart in why bond yields rise when central banks cut rates. Both halves of that sentence matter here. The equity run means most portfolios have drifted overweight risk. The yield level means the bond sleeve you would be buying back into is offering more than it has in years.
That is not a market call, and this article does not make one. It is an observation that the conditions which make people least willing to rebalance are the conditions that make rebalancing most worth doing.
Frequently asked questions
Does rebalancing improve returns?
Not reliably. In some periods selling winners and buying losers adds return, in others it subtracts, and over long samples the effect is small and depends heavily on the window. Rebalancing keeps risk near target. Treat any return effect as a rounding error rather than the reason to do it.
Should I rebalance during a crash?
That is precisely when the rule earns its keep, because it means buying what just fell. If your threshold has been breached, act on it. The reason to fix the rule in advance is that you will not want to when the moment comes.
Is quarterly better than annual?
No, on the evidence. Quarterly rebalancing of a 60/40 has not produced meaningfully better long-run risk or return than annual, and it trades more. If you want tighter control, use a drift threshold rather than a shorter calendar.
What if I have several accounts?
Rebalance the whole allocation as one portfolio, and do the trading inside the account where it costs least. Looking at each account separately produces trades that are unnecessary at the level that matters.
Does this change with age?
The target changes with age; the mechanism does not. As the horizon shortens, the equity weight should come down, and that is a deliberate change to the target rather than a rebalance. Our risk profile framework treats horizon as a ceiling for exactly this reason.
Related reading
- Is 60/40 still alive? — what the target should be before you maintain it.
- What type of investor are you? — the assessment that sets the target in the first place.
- Why bond yields rise when central banks cut rates — what the bond sleeve is currently paying you.
- Lump sum versus dollar-cost averaging — how to put new money in once the target is set.
- How to manage investment risk — where rebalancing sits among the decisions that actually matter.
Sources and assumptions. Equity and bond returns are the S&P 500 and the Bloomberg US Aggregate Bond Index over the periods stated. Drift figures are arithmetic applied to a portfolio starting at 60/40 with no contributions or withdrawals. Drawdown figures apply the roughly 55% peak-to-trough decline in global equities between October 2007 and March 2009 to each equity weight, holding the bond sleeve flat. Rebalancing frequency findings are from Vanguard’s published research on portfolio rebalancing, including its analysis of a 60/40 portfolio from 1926 and its threshold work for target-date funds. Figures were current at the date of publication and are not updated as markets move.
About this publication. This article is general information about portfolio maintenance and is not personalised investment advice: it does not take into account the financial situation, objectives or risk tolerance of any individual reader, and the same text is distributed to all readers. It is not a recommendation to buy or sell any financial instrument. Allocation figures illustrate how the arithmetic works and are not proposed portfolios. Tax treatment depends on your jurisdiction and your circumstances. Past performance does not indicate future results, and capital is at risk. AssetWhisper is not a registered investment adviser. Full disclaimer.
