Chapter 10: Rebalancing, the only free discipline
Run $10,000 at a 60/40 mix through five years like the ones loaded into this chapter's calculator: stocks return +25%, +15%, −20%, +30%, and +10% while bonds earn a steady 3% a year. Touch nothing and you finish with $14,504, which is a fine result. You also finish with a portfolio that is 68.0% stocks. You chose 60 back in Chapter 2 by reading the worst-year column until you found a loss you could hold. Nobody chose 68. The market votes every trading day, and five years of votes quietly rewrote the one big decision while you were busy living your life.
Rebalancing is the maintenance rule that votes back: selling a little of whatever grew past its target and buying a little of whatever fell behind, until the mix sits where you set it. It costs almost nothing to run, it requires no forecast, and it is the only discipline in this guide that works better the less you think about it. This chapter covers what the evidence says rebalancing is for, how often it deserves your attention, why new contributions should do most of the work in a taxable account, and why the rule has to be written down long before the day it is needed.
How a 60/40 becomes a 68/32
Stocks and bonds grow at different speeds, so their shares of the pot drift apart on their own. In four of the calculator's five years, stocks outgrew the steady 3% bond return by gaps between 7 and 27 points, and every gap handed stocks a slightly larger share of the total. The one losing year shoved the other way, hard, and still lost the tug of war. Drift is not a flaw in this particular path. Any stretch where one asset outruns the other produces it, and across most multi-year stretches since 1926, stocks have been the faster one.
The problem is not the extra money. The problem is that a 68/32 portfolio is a different machine from the one you stress-tested. On Vanguard's published model-allocation data (1926–2024), a 60/40's worst year was −26.6% and an 80/20's was −34.9%. Drifting from 60 to 68 carries you roughly two-fifths of the distance between those two losses, without a single decision being made. Plan line nine, your worst-year dollar line, was computed at 60/40. The portfolio that line described no longer exists.
The calculator runs that exact five-year path two ways. "Never" ends at 68.0% stocks and $14,504. The default 5-point band, which steps in whenever the stock weight wanders 5 points from target, triggers 3 times along the way and ends at 61.6% stocks and $14,668. Drag the returns around and try other bands; the pattern that survives every variation is the one this chapter is built on.
What rebalancing actually buys
Look at the two ending balances first: $14,668 against $14,504, a gap of $164 on $10,000 after five years. If rebalancing were a return machine, it would be a disappointing one, and other paths in the calculator will hand the money win to "Never" instead. Vanguard's research lineage on this question (2015 through 2024) reaches the same verdict at scale: rebalanced and unrebalanced portfolios do not differ in any dependable way on raw return, because the discipline sometimes trims a winner that keeps winning and sometimes buys a laggard that keeps lagging.
Now look at the two ending mixes: 61.6% stocks against 68.0%. That difference is the actual product. Rebalancing is risk control. It keeps the size of your next bad year close to the size you agreed to in writing, year after year, no matter what the market does to your weights. Vanguard's research frames the practice exactly this way, as managing risk rather than enhancing return, and the framing changes what counts as success. A rebalance that costs you some upside in a long bull market did not fail; it paid the premium on the insurance you asked for in Chapter 9.
How often is often enough?
Less often than almost everyone assumes. Vanguard compared monthly, quarterly, and annual rebalancing over long horizons and found risk-adjusted results that are not meaningfully different, while costs in trades, taxes, and attention rise with frequency. The practical answer comes in two flavors. The calendar flavor: pick one date a year and rebalance on it. The threshold flavor: act only when a weight drifts about 5 points from target, which on the calculator's path meant acting 3 times in five years. Vanguard's December 2024 follow-up found that threshold-based rebalancing slightly beat calendar-based for target-date funds, so the band carries a small edge for anyone willing to glance at their weights a few times a year. Either rule works, and the gap between them is tiny next to the gap between having a rule and having a mood. Pick the one you will actually follow and write it on the plan.
New money first, especially in taxable
Inside a 401(k) or an IRA, a rebalancing sale carries no tax bill, so sell whatever the band says to sell. In a taxable account, selling a winner realizes capital gains, which means the band can mail you an invoice. The cheaper move, and Vanguard's standing advice, is to rebalance with new cash first: aim every fresh contribution, plus any dividends you are not spending, at the underweight asset until the weights heal. A portfolio receiving steady contributions can live inside a 5-point band for years without one taxable sale. The full capital-gains arithmetic, including when a sale is worth making anyway, belongs to the Finvest Tax Playbook. The portable version is an order of operations: new money first, tax-sheltered sales second, taxable sales last and reluctantly.
The moment will argue
Notice what the band demanded in the calculator's third year. Stocks had just lost 20%, and the rule's answer was to buy more of them with bond money. That is rebalancing at nearly every trigger: a purchase order for the thing that just embarrassed itself, signed in advance by a calmer version of you. When the real moment arrives there will be headlines explaining why this time the loser stays down, and a reasonable-sounding voice suggesting you wait a few weeks for clarity. The written rule exists precisely because the moment will argue, and the moment is persuasive. Chapter 9 measured your worst-year line on a day when nothing was burning. This chapter's rule is how that measurement keeps its authority on the day something is.
Dev left the ETFs & Funds Guide with two bands he had never actually written down: 5 points on his 80/20 stock-bond split, and 5 points on the international share of his stocks, which targets 30%. Tonight they become plan text, with the order of operations underneath: new 401(k) money aims at whatever is underweight, IRA sales come second, taxable sales last. He tests the rule against an imaginary crash, notices how badly he wants an exception clause, and adds one sentence instead: "A band is a tripwire I set on a quiet day, and it outranks anything I think on a loud one."
Where rebalancing goes wrong
The first wrong turn is overdoing it: checking daily, acting monthly, mistaking activity for control. The frequency evidence says the extra effort buys nothing measurable, and in a taxable account the extra trades cost real money. The second is smuggling forecasts back in, skipping a trigger because stocks "have momentum" or jumping one early because a correction "feels close." The band holds no opinion about the future; firing without one is its entire value. The third is rebalancing taxable accounts with sales when contributions could have done the job silently. The fourth is rebalancing a portfolio that was never written down, which is not rebalancing at all, just trading with extra steps.
Pick one trigger and write it down: a 5-point band or a single annual date. When it trips, act without consulting your feelings or the news, and let new money do the work in taxable accounts. The rule gets set on a quiet day because it will be executed on a loud one.
Line ten, after nine lines that chose your goals, your mix, and your worst-year dollar line: "I rebalance when a weight drifts 5 points from target or at my annual review, whichever comes first, using new money first in taxable accounts." Two chapters remain before the page is full.
Key takeaways
- Drift is automatic: on the calculator's five-year path, an untouched 60/40 ends at 68.0% stocks and $14,504, a riskier portfolio nobody chose.
- The 5-point band ends that same path at 61.6% stocks and $14,668. The $164 gap is noise; the 6.4-point risk gap is the product.
- Per Vanguard's research lineage, rebalancing is risk control, not return enhancement, and monthly, quarterly, and annual frequencies produce risk-adjusted results that are not meaningfully different. Annual dates or 5-point bands are the practical answer, and the December 2024 follow-up gives thresholds a slight edge for target-date funds.
- In taxable accounts, rebalance with new money first; selling winners triggers capital gains the Tax Playbook can help you price.
- Every trigger asks you to buy what just lost. Write the rule while calm, because the moment will argue.
Sources: Vanguard rebalancing research (December 2024) · Vanguard model-allocation data · Finvest Tax Playbook · Finvest ETFs & Funds Guide