How often should investors rebalance a portfolio? Vanguard's research answer is once a year, or once a year with a tolerance band of 1 to 2 percentage points, because monthly and quarterly schedules add trading costs without adding risk control, while waiting two years lets risk drift too far [1]. The 2026 market shows how drift happens in real time: a 60/40 mix of the S&P 500 (SPX) and 3-month Treasury bills started the year at 60% stocks, slipped to 58.6% stocks at the March 31 low, and stood at 61.7% stocks on June 30 after the index rallied to 7,499.36 [2]. Over long stretches the drift is far larger. Vanguard notes that a 60/40 portfolio set at the end of 2003 and never rebalanced would have been 80% stocks by the end of 2022 [3].
What rebalancing is for
Rebalancing is a risk tool, not a return tool. Vanguard's January 2026 note states that the primary function is to keep a portfolio aligned with the investor's tolerance for risk, not to maximize returns [3]. Selling the asset that has risen and buying the one that has fallen also forces a contrarian discipline that most investors find hard to follow by instinct.
The cost of never rebalancing shows up as unintended risk. A 60/40 investor who ended 2003 with a balanced portfolio and left it alone was effectively an 80/20 investor by the end of 2022, and carried 80/20 losses through the 2022 bear market when the S&P 500 fell 18.04% and 10-year Treasuries fell 17.83% [3] [4].
How often should investors rebalance a portfolio?
Vanguard's October 2022 paper, Rational Rebalancing, tested calendar rules (monthly, quarterly, semiannual, annual), threshold rules (rebalance when an asset drifts 1% to 15% from target) and combinations of the two, using simulated returns and transaction costs [1]. The finding was that the optimal method is neither too frequent, such as monthly or quarterly, nor too infrequent, such as every two years. Annual rebalancing, or annual with a 1% or 2% threshold, was optimal for investors who do not harvest tax losses or track a benchmark tightly [1].
Why more frequent is not better
Each rebalancing trade pays a spread and, in taxable accounts, may realize gains. Frequent rebalancing also cuts short the momentum that markets often show over periods of a few months. Vanguard's January 2026 note adds that transaction costs rise during volatile periods, which makes frequent rebalancing less efficient exactly when portfolios drift the most [3]. Partial rebalancing, or directing new contributions and dividends to the underweight asset, achieves most of the risk control at a fraction of the cost [3].
By how much: the case for tolerance bands
A tolerance band says: do nothing until an asset class is more than X percentage points away from target, then trade back. Vanguard found no material difference between a 1% and a 2% band when combined with an annual review [1]. Wider bands, such as 5 percentage points on a 60% stock target, are common among advisors because they cut trading further; the trade-off is a somewhat larger risk drift between trades.
- Calendar only: simple, predictable, and adequate for most investors when done annually [1].
- Threshold only: reacts to markets rather than dates, but requires monitoring and can trigger many trades in volatile years.
- Calendar plus threshold: check once a year, trade only if the drift exceeds the band. This is the combination Vanguard rates as most efficient [1].
- Cash-flow rebalancing: send contributions, dividends and interest to the underweight asset; this avoids selling anything [3].
What rebalancing did in the 2026 market
The first half of 2026 was a useful stress test because it contained both a drawdown and a recovery. The S&P 500 closed 2025 at 6,845.50, fell to 6,528.52 on March 31 (and to 6,343.72 on March 30, about 9% below its late-January high), then rallied to 7,499.36 by June 30 [2]. Three-month Treasury bills yielded between 3.57% and 3.74% over the same period, so the cash side of a 60/40 portfolio earned a steady return [5].
| 60/40 S&P 500 and T-bills, start value 100 on Dec 31, 2025 | Value Mar 31 | Stock weight Mar 31 | Value Jun 30 | Stock weight Jun 30 |
|---|---|---|---|---|
| Never rebalanced | 97.58 | 58.6% | 106.46 | 61.7% |
| Rebalanced once at Mar 31 | 97.58 | 60.0% | 106.65 | 63.1% |
| Rebalanced monthly | 97.58 | 60.0% | 106.59 | 60.0% |
Calculations use S&P 500 price closes and 3-month bill yields from FRED [2] [5]. The results illustrate three lessons. First, drift over six months was modest, less than 2 percentage points, so an investor who did nothing was not far off target. Second, the single threshold trade at the March low added about 0.2 points of value because it bought stocks before the rebound. Third, monthly rebalancing added trades but slightly less value than the single well-timed trade, which is consistent with the research finding that frequency alone does not help [1]. Investors who want to monitor drift day to day can use price history for index funds and ETFs on DataPorium's ETF page [6].
Bonds, cash and the 2022 lesson
Rebalancing into bonds only reduces risk if bonds behave differently from stocks. In 2022 they did not: both fell by about 18% [4]. In 2026, with 3-month bills yielding 3.74% as of June 30, short-term Treasuries offer a rebalancing destination that carries little duration risk [5]. The trade-off is that bills provide no upside if yields fall sharply in a recession, which is when long bonds have historically done their best work.
Rebalancing once a year with a small tolerance band captures nearly all of the risk control at the lowest cost.
Key takeaways
- Vanguard's research supports annual rebalancing, or annual with a 1% to 2% threshold, over monthly, quarterly or two-year schedules [1].
- A 60/40 portfolio left alone from end-2003 became 80/20 by end-2022; drift is small in one year and large over two decades [3].
- In the first half of 2026, an unrebalanced 60/40 S&P 500 and T-bill mix drifted only to 61.7% stocks, and a single rebalance at the March low added about 0.2 points [2] [5].
- The purpose is risk alignment, not return; the 2022 losses in both stocks and bonds show why the destination asset matters [3] [4].
Frequently asked questions
How often should I rebalance my portfolio?
Vanguard's research finds that once a year, or once a year with a 1% to 2% tolerance band, is optimal for most investors who are not harvesting tax losses [1]. Monthly and quarterly rebalancing adds costs without adding risk control.
What is a rebalancing threshold?
A threshold is the amount an asset class may drift from its target before a trade is made, for example 5 percentage points on a 60% stock target. Vanguard found no material difference between 1% and 2% bands when combined with an annual check [1].
Did rebalancing help in 2026?
Modestly. A 60/40 mix of the S&P 500 and Treasury bills rebalanced at the March 31, 2026 low finished June at 106.65 versus 106.46 for a portfolio left alone, using FRED price and yield data [2] [5].
Does rebalancing increase returns?
Not reliably. Its main purpose is to keep risk in line with the investor's tolerance [3]. It can add value when it buys after a fall, as in March 2026, and it can subtract value in long one-directional trends.
Sources & References
- [1] Vanguard research: Rational rebalancing, an analytical approach to multiasset portfolio rebalancing (October 2022)
- [2] FRED: S&P 500 (SP500), daily close
- [3] Vanguard: Why, how and when multi-asset investors should rebalance (January 2026)
- [4] NYU Stern (Damodaran): Historical Returns on Stocks, Bonds and Bills, 1928 to 2025
- [5] FRED: 3-Month Treasury Bill Secondary Market Rate (DTB3)
- [6] DataPorium ETF Analytics
